CN117010938A - Order quantity optimizing system, order quantity optimizing method, and recording medium - Google Patents

Abstract

The invention provides an order quantity optimization system, an order quantity optimization method and a recording medium. The order quantity optimization system has: a data collection unit that collects historical data of business data related to commodities; a demand forecasting unit that uses multiple prediction algorithms for each commodity to predict future demand based on the historical data; and prediction algorithm selection A unit that, for each commodity, calculates the prediction accuracy of each of the plurality of prediction algorithms using the historical demand prediction values of the plurality of prediction algorithms in the past given period and the actual demand value in the given period. , select the prediction algorithm with the highest prediction accuracy, and obtain the demand prediction value of the prediction algorithm with the highest prediction accuracy regarding the future demand from the demand prediction unit. As a result, order business can be completed quickly and accurately.

Description

Order volume optimization system, order volume optimization method and recording medium

Technical field

The invention relates to an order quantity optimization system, an order quantity optimization method and a recording medium.

Background technique

In various businesses such as manufacturing and product sales, it is necessary to purchase a variety of materials, equipment, products and other commodities. In order to match the purchased order volume with demand and inventory conditions, order plans need to be formulated reasonably. In the past, skilled workers on site made order plans based on their own experience, including selecting product types, determining order quantities, specifying delivery dates, etc. This not only required a lot of manpower and time, but also involved order volume and demand. The situation and inventory situation do not match, which affects production and sales. Therefore, it is hoped that the business of making order plans can be digitized and order plans can be made automatically and accurately.

Patent Document 1 proposes an intelligent replenishment system based on supply chain demand. The system consists of a demand forecast module, a business analysis module and an intelligent replenishment module. The demand forecast module is used to predict demand for each store and each product. The business analysis module is used to classify demand according to the relevant information of the product and related sales targets, and in accordance with the results of demand forecasting. The intelligent replenishment module is used to use the results of the demand forecast module and business analysis module to automatically generate inventory replenishment plans corresponding to demand classification.

However, the intelligent replenishment system of Patent Document 1 has the following problems.

First, in terms of demand forecasting, Patent Document 1 only briefly mentions some forecasting algorithms and does not describe how to use these forecasting algorithms for forecasting. In actual business, different commodities have different characteristics. For example, the demand for some commodities changes relatively steadily, while the demand for some commodities fluctuates greatly. For commodities with large demand changes, even if machine learning or AI models are used for prediction, the prediction accuracy is often relatively low. In other words, although there are many prediction algorithms, there is no prediction algorithm suitable for all commodities. Therefore, there is a problem that it is difficult to select an appropriate prediction algorithm for each commodity.

Secondly, in terms of inventory replenishment, Patent Document 1 adopts a fixed replenishment strategy formulated in advance for the product. In actual business, in most cases the inventory required for each commodity changes in real time. Therefore, if replenishment is carried out according to a fixed replenishment strategy, there is a problem that the actual replenishment requirements cannot be met. In addition, usually, the replenishment plan cannot be implemented directly as an order plan because there are capacity constraints on the supplier. Therefore, if the replenishment plan does not match the supplier's production capacity, the replenishment plan needs to be adjusted to formulate the final order plan. Patent Document 1 only relates to the replenishment plan and does not relate to the order plan, and there is a problem that the replenishment plan of each product needs to be appropriately adjusted.

existing technical documents

Patent document 1: CN110516998 A

Contents of the invention

The present invention is proposed to solve the above problems, and its purpose is to provide an order volume optimization system, an order volume optimization method, and a recording medium that can quickly and accurately implement order business.

Means used to solve problems

In order to achieve the above object, according to the first aspect of the present invention, an order volume optimization system is provided, which is provided with: a data collection unit to collect historical data of business data related to commodities; a demand forecasting unit to predict and predict business data based on the historical data. each commodity utilizes multiple forecasting algorithms to predict future demand; and a forecasting algorithm selection unit that, for each commodity, forecasts historical values of demand using the multiple forecasting algorithms for a given period in the past and demand within the given period. Actual value, calculate the prediction accuracy of each prediction algorithm among the plurality of prediction algorithms, select the prediction algorithm with the highest prediction accuracy, and obtain the prediction algorithm with the highest prediction accuracy from the demand prediction unit about the future demand demand forecast value.

According to a second aspect of the present invention, an order quantity optimization system is provided, which is provided with: a data collection unit that collects historical data of business data related to commodities; a demand forecasting unit that predicts the future of each commodity based on the historical data. Forecasting demand; and an inventory standard setting unit that, for each commodity, obtains a demand characteristic index indicating the importance of the commodity based on the historical data, and based on the obtained demand characteristic index and the demand prediction unit obtained from the demand prediction unit Demand forecast values set inventory standards at multiple levels.

According to a third aspect of the present invention, an order volume optimization system is provided, which is provided with: a data collection unit that collects historical data of business data related to commodities; a demand forecasting unit that predicts the future of each commodity based on the historical data. Demand is forecasted; an inventory simulation unit is configured to, for each commodity, simulate changes in inventory within a given period in the future based on the current inventory, remaining orders, and demand forecast values obtained from the demand forecasting unit; and The replenishment plan generation unit determines, for each commodity, the replenishment priority and quantity of the commodity that currently need to be replenished based on the inventory of the final unit period in the future given period obtained from the inventory simulation unit. The replenishment amount of each stall generates a replenishment plan including the replenishment priority of all products as replenishment objects and the replenishment amounts of multiple stalls.

According to a fourth aspect of the present invention, an order volume optimization system is provided, including: a data collection unit that collects historical data of business data related to commodities; a demand forecasting unit that predicts the future of each commodity based on the historical data. Demand quantity is forecasted; a replenishment plan generation unit generates, for each commodity, a replenishment plan including the replenishment quantity of all commodities as replenishment objects based on the prediction results of the demand prediction unit; and an order plan generation unit, The replenishment quantity in the replenishment plan is adjusted to match the supplier's production capacity, the adjusted replenishment quantity is used as the order quantity, and an order plan including the order quantity of all commodities is generated.

According to the fifth aspect of the present invention, an order volume optimization method is provided, including: a data collection step, which collects historical data of business data related to commodities; a demand forecasting step, which uses multiple data for each commodity based on the historical data. A forecasting algorithm is used to predict future demand; and a forecasting algorithm selection step, for each commodity, uses the historical demand forecast values of the multiple forecasting algorithms in the past given period and the actual demand value in the given period to calculate the The prediction accuracy of each prediction algorithm among the plurality of prediction algorithms is determined, the prediction algorithm with the highest prediction accuracy is selected, and the demand prediction value of the prediction algorithm with the highest prediction accuracy in the demand prediction step regarding the future demand is obtained. .

According to the sixth aspect of the present invention, an order volume optimization method is provided, including: a data collection step to collect historical data of business data related to commodities; a demand forecasting step to predict the future of each commodity based on the historical data. Forecasting demand; and an inventory standard setting step, for each commodity, obtaining a demand characteristic index indicating the importance of the commodity based on the historical data, and based on the obtained demand characteristic index and the demand in the demand forecasting step Forecast values, setting inventory standards at multiple levels.

According to the seventh aspect of the present invention, an order quantity optimization method is provided, including: a data collection step to collect historical data of business data related to commodities; a demand forecasting step to predict the future of each commodity based on the historical data. Forecasting demand; an inventory simulation step, for each commodity, simulating changes in inventory within a given period in the future based on the current inventory, remaining orders, and the demand forecast value in the demand forecasting step; and supplementing In the inventory plan generation step, for each commodity, according to the inventory amount in the final unit period in the future given period in the inventory simulation step, determine the replenishment priority and multiple slots that the commodity currently needs to replenish. Based on the replenishment quantity of each position, a replenishment plan is generated including the replenishment priority of all products as replenishment objects and the replenishment quantity of multiple levels.

According to an eighth aspect of the present invention, an order volume optimization method is provided, including: a data collection step, which collects historical data of business data related to commodities; and a demand forecasting step, which predicts the future of each commodity based on the historical data. Forecasting the demand quantity; a replenishment plan generating step, for each commodity, generating a replenishment plan including the replenishment quantities of all commodities as replenishment objects according to the prediction results of the demand forecasting step; and an order plan generating step, The replenishment quantity in the replenishment plan is adjusted to match the supplier's production capacity, the adjusted replenishment quantity is used as the order quantity, and an order plan including the order quantity of all commodities is generated.

According to a ninth aspect of the present invention, there is provided a computer-readable recording medium storing a program, wherein the program is used to cause a computer to execute the above order quantity optimization method.

Invention effect

According to the order volume optimization system, the order volume optimization method and the recording medium of the present invention, the order business can be automatically executed without relying on the experience of a skilled person, and the order business can be realized quickly and accurately. This can improve the efficiency and accuracy of inventory management. Since inventory can be properly managed, there will be no overstocking or shortage of stock. Therefore, risks such as difficulties in capital turnover caused by inventory backlog and production shutdown caused by shortage can be prevented.

Description of the drawings

FIG. 1 is a schematic diagram showing an example of an application object to which the order quantity optimization system and method of the present invention are applied.

FIG. 2 is a block diagram showing the functional structure of the order quantity optimization system according to the first embodiment of the present invention.

FIG. 3 is a flowchart showing the order quantity optimization method according to the first embodiment of the present invention.

FIG. 4 is a block diagram showing the functional structure of the order quantity optimization system according to the second embodiment of the present invention.

FIG. 5 is a diagram showing classification based on the ABC analysis method according to the second embodiment of the present invention.

6 is a block diagram showing the functional structure of the order quantity optimization system according to the third embodiment of the present invention.

7 is a block diagram showing the functional structure of the order quantity optimization system according to the fourth embodiment of the present invention.

FIG. 8 is a block diagram showing the functional structure of the order quantity optimization system according to the fifth embodiment of the present invention.

FIG. 9 is a flowchart showing an order quantity optimization method according to the fifth embodiment of the present invention.

FIG. 10 is a block diagram showing the functional structure of the order quantity optimization system according to the sixth embodiment of the present invention.

FIG. 11 is a flowchart showing an order quantity optimization method according to the sixth embodiment of the present invention.

FIG. 12 is a block diagram showing the functional structure of the order quantity optimization system according to the seventh embodiment of the present invention.

FIG. 13 is a flowchart showing an order quantity optimization method according to the seventh embodiment of the present invention.

Detailed ways

The present invention will be described in more detail below with reference to the drawings, embodiments and specific examples. However, the following description is merely an example to facilitate understanding of the present invention, and is not intended to limit the scope of the present invention. In the specific implementation, the components and units of the device and system can be changed, deleted or added according to the actual situation, and the steps of the method can be changed, deleted, added or changed in order according to the actual situation. The same or equivalent components, steps, etc. shown in the respective drawings are denoted by the same symbols, and repeated descriptions are appropriately omitted.

First, the application objects to which the order volume optimization system and method of the present invention can be applied are explained. FIG. 1 is a schematic diagram showing an example of an application object to which the order quantity optimization system and method of the present invention are applied. In FIG. 1 , three factories, Factory 1, Factory 2, and Factory 3, and two suppliers, Supplier 1 and Supplier 2, are shown. However, the number of factories and suppliers is not limited to this. .

The order quantity optimization system and method of the present invention can be applied to, for example, factory 1, factory 2, and factory 3. Each factory needs to purchase goods from each supplier when carrying out production activities. In addition, for example, when supplier 1 needs to purchase goods from supplier 2, the order quantity optimization system and method of the present invention can also be applied to supplier 1. That is to say, the order quantity optimization system and method of the present invention can be applied to all objects that need to purchase goods.

In the present invention, "commodity" represents the object of purchase, which may be materials used for manufacturing, or production equipment, etc., and "order quantity" represents the purchase quantity of the commodity in the order.

By applying the order volume optimization system and method of the present invention, the order business can be automatically executed without relying on the experience of a skilled person, and the order business can be realized quickly and accurately. This can improve the efficiency and accuracy of inventory management. Since inventory can be properly managed, there will be no overstocking or shortage of stock. Therefore, risks such as difficulties in capital turnover caused by inventory backlog and production shutdown caused by shortage can be prevented.

(first embodiment)

FIG. 2 is a block diagram showing the functional structure of the order quantity optimization system 100 according to the first embodiment of the present invention. As shown in FIG. 2 , the order quantity optimization system 100 includes a data collection unit 110 , a demand prediction unit 120 , a prediction algorithm selection unit 130 and a memory 140 , which are connected to each other through a bus 150 .

<data collection unit>

The data collection unit 110 collects historical data of business data related to merchandise. As shown in FIG. 2 , the data collection unit 110 can collect historical data from the memory 140 , but is not limited thereto. When the historical data is stored in a memory external to the order volume optimization system 100 , the historical data is obtained from the external memory. That’s it.

Historical data of business data related to commodities includes, for example, historical data of incoming and outgoing records, historical data of inventory records, historical data of order records, etc.

Table 1 shows an example of historical data of inbound and outbound records. In the example shown in Table 1, the historical data of the inbound and outbound record includes the date of inbound and outbound warehouse, warehouse ID, product number, classification indicating whether it is inbound or outbound, the quantity and amount of the corresponding product, order number, and the order number. The line number marked when multiple products are included. Of course, other data can also be included as needed. Usually, when certain commodities need to be used, corresponding outbound orders will be generated, including information such as product numbers, quantities, and amounts of these commodities. The management system records outbound shipments based on the information in the outbound orders. In addition, when the goods purchased from suppliers arrive, there is also a corresponding invoice, including the product number, quantity and other information of these goods. The management system performs warehousing records based on the information in the invoice.

[Table 1]

date Warehouse ID Product Number Classification quantity Amount order number Line number 2018/1/1 LOC01 PT001 Warehouse 36 36 PO-001 1 2018/1/1 LOC01 PT001 out of warehouse twenty two twenty two 2018/1/3 LOC01 PT001 Warehouse 33 33 PO-002 1 … … … … … …

An example of historical data for inventory records is shown in Table 2. In the example shown in Table 2, the historical data of the inventory record includes date, warehouse ID, item number, quantity and amount of the item. Of course, other data can be included as needed. Usually, inventory of goods is carried out according to a given cycle, such as once a day, every week or every month, and the generated information includes the date of the count, the warehouse ID of the stocktake, the number and quantity of each commodity stored in each warehouse. Wait for the inventory results, and the management system will perform inventory records based on the information in the inventory results. In addition, the management system can also record inventory based on the results of settling the incoming and outgoing flows according to a given period.

[Table 2]

date Warehouse ID Product Number quantity Amount 2018/1/1 LOC01 PT001 40 40 2018/1/1 LOC01 PT001 35 35 2018/1/3 LOC01 PT001 56 56 … … … … …

An example of historical data for order records is shown in Table 3. In the example shown in Table 3, the historical data of the order record includes warehouse ID, order number, line number marked when the order includes multiple items, supplier number, item number, order quantity and amount, order generation date, Order confirmation date, scheduled arrival date, actual arrival date, latest arrival date, receipt quantity and amount, and of course other data can be included as needed. Based on the historical data of order records, you can learn information such as the number of past orders, which orders have been completed, and which orders have not yet arrived.

[table 3]

<Demand Forecasting Unit>

The demand prediction unit 120 uses a plurality of prediction algorithms for each commodity to predict future demand based on historical data of business data related to the commodity collected by the data collection unit 110 .

The prediction algorithm can be a well-known appropriate algorithm. In this embodiment, four time series analysis algorithms are used, including weighted moving average (hereinafter also referred to as WMA) and exponential smoothing (hereinafter also referred to as WMA). (EMA), Holt-Winters method (Holt-Winters method, also referred to as HoltWinters below), ARIMA method (Autoregressive integrated moving Average, also referred to as ARIMA below).

In the weighted moving average algorithm, different weights are given to each time series data to calculate the average. In the exponential smoothing method, the newer data among the observation values are given a larger weight, and the earlier data are reduced by an exponential function in a weighted manner, and a moving average is calculated. In the Holt-Winter method, trends and seasonal fluctuations are added to the time series changes in the exponential smoothing method, and the superposition of the respective exponential smoothing results is calculated as the expected value. The model using the ARIMA method combines three models: autoregressive model (AR model), moving average model (MA model), and integrated model (I model). It is also called an integrated moving average autoregressive model. In time series data When there is a cyclical change or trend change, predictions that take into account the impact of the cyclical change or trend change can be made.

The demand forecasting unit 120 at least needs to predict the demand within the purchase cycle of the product. The procurement cycle refers to the period from the placement of the order to the arrival of the goods, and can be calculated by adding the order period plus the supplier's delivery period. For example, when the order period is one month and the delivery period is three months, the procurement cycle is four months, and the demand forecasting unit 120 needs to predict the demand for the next four months.

<Prediction algorithm selection unit>

The prediction algorithm selection unit 130 is configured to automatically select an optimal prediction algorithm for each commodity from a plurality of prediction algorithms used by the demand prediction unit 120, and output prediction results using the optimal prediction algorithm.

To this end, the forecast algorithm selection unit 130 calculates the forecast accuracy of each forecast algorithm for each commodity by using the demand forecast history values of multiple forecast algorithms in the past given period and the actual demand value in the given period, and selects Among them, the prediction algorithm with the highest prediction accuracy is used as the optimal prediction algorithm for the commodity, and then the prediction result of the optimal prediction algorithm regarding the future demand for the commodity is obtained from the demand prediction unit 120, and the prediction result is used as the subsequent order for determining the commodity. Demand forecast values used in the quantity process.

When calculating the prediction accuracy of a prediction algorithm, multiple well-known prediction accuracy evaluation indexes (hereinafter also referred to as accuracy indexes) can be used. In this implementation, three accuracy indicators, MAE (Mean Absolute Error), RMSE (Root Mean Square Error), and MAPE (Mean Absolute Percent Error), are used. The calculation methods are as shown in Equations (1) to (3) respectively. ,

In formulas (1) to (3), represents the historical demand forecast value of the i-th unit period in the past given period, y _i represents the actual demand value of the i-th unit period in the past given period, n represents the number of unit periods in the past given period, each unit There is 1 demand forecast historical value and 1 demand actual value during the period. Here, let the past given period be one procurement cycle before the current time point. For example, the procurement cycle is four months and the unit period is one month. In this case, n is 4.

Among the three accuracy indicators of MAE, RMSE, and MAPE, MAE and RMSE can reflect the error of the predicted value. The smaller the values of MAE and RMSE, the smaller the error of the predicted value. MAPE can reflect the error size of the model. The accuracy of the model can be obtained by subtracting the value of MAPE from 100%. Therefore, the smaller the value of MAPE, the smaller the error and the higher the accuracy of the model.

The following describes an example of a process in which the prediction algorithm selection unit 130 selects an optimal prediction algorithm for the target product, taking the product with the product number "PT001" as the target product.

Assume that the current time point is August 1, 2021, the procurement cycle is 4 months, the unit period is 1 month, and the 1 procurement cycle before the current time point is from April to July 2021. At the current time point, it can be obtained The actual demand value of the target commodity during the procurement cycle as shown in Table 4 represents the actual consumption of the target commodity during this period.

[Table 4]

Warehouse ID Product Number April May June July LOC01 PT001 253 235 256 251

Assume that the demand forecasting unit 120 uses four forecasting algorithms WMA, EMA, HoltWinters, and ARIMA to predict the demand for the target commodity from April to July 2021 on April 1, 2021, and obtains the forecast results shown in Table 5. .

[table 5]

execution date Prediction algorithm Warehouse ID Product Number April May June July 20210401 WMA LOC01 PT001 217 206 228 260 20210401 EMA LOC01 PT001 198 190 182 235 20210401 HoltWinters LOC01 PT001 237 278 269 230 20210401 ARIMA LOC01 PT001 243 278 284 262

As shown in Table 5, for each forecasting algorithm, the historical value of demand forecast in the past given period (April to July) is obtained, so that the historical value of demand forecast can be used as well as the demand forecast value in this period as shown in Table 4. Use the actual value of demand to evaluate the prediction accuracy of each prediction algorithm and select the optimal prediction algorithm. The specific process includes the following steps (a), (b), and (c).

(a) Calculate the accuracy index of each prediction algorithm

According to the calculation method shown in the above formulas (1) to (3), the actual demand value ( _yi ) in Table 4 and the demand forecast historical value in Table 5 are used The three accuracy indicators of MAE, RMSE, and MAPE are calculated for each prediction algorithm, and the calculation results are listed in Table 6.

[Table 6]

execution date Prediction algorithm Warehouse ID Product Number MAE RMSE MAPE 20210401 WMA LOC01 PT001 25.50 27.40 0.10 20210401 EMA LOC01 PT001 47.50 51.92 0.19 20210401 HoltWinters LOC01 PT001 23.25 26.05 0.10 20210401 ARIMA LOC01 PT001 23.00 26.71 0.09

Shown in Table 6 are the accuracy indicators calculated for the forecast results on April 1, 2021. Similarly, the given period in the past can be set as February to May, March to June, etc. in 2021, and the accuracy indicators of the prediction results at different times can be obtained.

(b) Calculate the weighted moving average of the accuracy indicator

When a prediction algorithm is evaluated using only one calculation result of the accuracy index, it may not be possible to accurately evaluate it due to the presence of chance factors. Therefore, in order to improve the accuracy of the evaluation, it is preferable to use the calculation results of the accuracy index multiple times for evaluation. For example, the calculation results of the accuracy index in the past three months, four months, or half a year can be used.

In this example, for the prediction algorithm WMA, the calculation results of the accuracy indicators for the last three months shown in Table 7 are used, that is, the calculation results for February 1, 2021, March 1, 2021, and April 2021 are used. The accuracy index calculated from the forecast results on the 1st of the month.

[Table 7]

execution date Prediction algorithm Warehouse ID Product Number MAE RMSE MAPE 20210201 WMA LOC01 PT001 15.80 18.98 0.08 20210301 WMA LOC01 PT001 22.78 23.52 0.12 20210401 WMA LOC01 PT001 25.50 27.40 0.10

Moreover, as shown in equation (4), for each accuracy index, the three calculation results are weighted moving average, and the weighted moving average AVG_MAE, AVG_RMSE, and AVG_MAPE of each accuracy index are obtained.

When calculating a weighted moving average of calculation results of multiple accuracy indexes, it is preferable to set the weight so that calculation results closer to the current time point have greater weight, thereby further improving the accuracy of the evaluation. For example, in equation (4), the calculation results of the accuracy index on April 1, 2021, March 1, 2021, and February 1, 2021 are given weights of 3, 2, and 1 respectively.

For the other three prediction algorithms EMA, HoltWinters, and ARIMA, the same calculation is performed to obtain the weighted moving averages AVG_MAE, AVG_RMSE, and AVG_MAPE of each accuracy indicator of each prediction algorithm.

(c) Select the optimal prediction algorithm

At the current point in time, sort the weighted moving averages of the accuracy indicators of each prediction algorithm obtained through step (b) in order from small to large. The smaller the weighted moving average of each accuracy indicator, the higher the prediction accuracy of the corresponding prediction algorithm. Therefore, the prediction algorithm with the highest prediction accuracy among the four prediction algorithms can be selected based on the weighted moving average as the target product. Optimal prediction algorithm.

When sorting, give priority to AVG_MAE, AVG_RMSE, and AVG_MAPE, so that the weighted moving averages AVG_MAE and AVG_RMSE, which reflect the error size of the prediction value, have higher priority, and make the weighting of MAPE, which reflects the accuracy of the prediction model, higher. The priority of the moving average AVG_MAPE is lower. For example, the priority is set to decrease in the order of AVG_MAE, AVG_RMSE, and AVG_MAPE. Therefore, when sorting, first sort according to the order of AVG_MAE from small to large. For prediction algorithms with the same AVG_MAE value, sort according to the order from small to large AVG_RMSE. For prediction algorithms with the same values of AVG_MAE and AVG_RMSE, sort according to the order of AVG_MAPE from small to large. Sort in order of largest. In this example, the sorting results shown in Table 8 are obtained.

[Table 8]

execution date Prediction algorithm AVG_MAE AVG_RMSE AVG_MAPE 20210801 WMA 22.98 24.70 0.10 20210801 EMA 24.73 25.82 0.08 20210801 HoltWinters 25.62 27.32 0.10 20210801 ARIMA 32.46 29.67 0.13

As shown in Table 8, among the four prediction algorithms, WMA has the smallest AVG_MAE, which means that for the target commodity, the prediction algorithm WMA has the highest prediction accuracy. Therefore, the prediction algorithm WMA is selected as the optimal prediction algorithm for the target commodity, and the optimal prediction algorithm is output. The prediction results of the optimal prediction algorithm.

Through the above steps (a), (b), and (c), the optimal prediction algorithm is selected for the target product with the product number "PT001". For other products, just perform the same steps to select the corresponding optimal prediction algorithm.

<Order volume optimization method>

Hereinafter, the order quantity optimization method executed by the order quantity optimization system 100 of this embodiment will be described using FIG. 3 . FIG. 3 is a flowchart showing the order quantity optimization method according to this embodiment.

As shown in Figure 3, the order quantity optimization method of this embodiment includes a data collection step S110, a demand prediction step S120, and a prediction algorithm selection step S130.

In step S110, historical data of business data related to the product is collected.

In step S120, multiple prediction algorithms are used for each commodity to predict future demand based on the historical data of the business data related to the commodity collected in step S110.

In step S130, for each commodity, the prediction of each prediction algorithm utilized in step S120 is calculated using the demand prediction history values of the multiple prediction algorithms in the past given period and the actual demand value in the given period. accuracy, and select the prediction algorithm with the highest prediction accuracy as the optimal prediction algorithm for the product, and then use the prediction result of the prediction algorithm with the highest prediction accuracy in step S110 as the demand forecast used in the subsequent process of determining the order quantity of the product. value.

According to the order quantity optimization system and order quantity optimization method of this embodiment, the optimal prediction algorithm is automatically selected for each commodity through the prediction algorithm selection unit or step, and the prediction result of the optimal prediction algorithm is used to determine the order quantity. , so the order business can be automatically executed without relying on the experience of a skilled person, and the order business can be realized quickly and accurately. This can improve the efficiency and accuracy of inventory management. Because inventory can be properly managed, there will be no overstocking or out-of-stock situations, so it can prevent risks such as difficulties in capital turnover caused by overstocking and production shutdowns caused by out-of-stocks.

(Second Embodiment)

FIG. 4 is a block diagram showing the functional structure of the order quantity optimization system 100A according to the second embodiment of the present invention. The difference between the order quantity optimization system 100A of this embodiment and the order quantity optimization system 100 of the first embodiment is that, as shown in FIG. 4 , the order quantity optimization system 100A also has an inventory standard setting unit 160. This difference will be used below. The point is centered.

In order to properly manage inventory, inventory standards such as safety stock need to be set. In Patent Document 1, fixed inventory standards are set based on actual consumption of products. However, in actual business, inventory standards need to be adjusted in real time according to different periods, demand conditions, etc. If fixed inventory standards are used as in Patent Document 1, inventory overstock or shortages are likely to occur.

In this embodiment, the inventory standard is dynamically set by the inventory standard setting unit 160 . Next, the inventory standard setting unit 160 will be described in detail.

<Inventory standard setting unit>

The inventory standard setting unit 160 is used to dynamically set corresponding inventory standards for each commodity.

The calculation formula of the inventory standard is shown in the following formulas (5) to (7).

Inventory quantity = usage quantity during the order period + safety stock (5)

Among them, the usage within the order period is the demand forecast value within the order period obtained by using the optimal forecasting algorithm, n is the number of sample data, which can be appropriately selected according to the actual situation, _Xi is the sample data, and μ is the sample data The average value of , the safety factor can be found from Table 9 according to the out-of-stock probability. The out-of-stock probability indicates the probability of a product being out-of-stock. In order to reduce the out-of-stock probability, the safety factor needs to be increased to increase the safety inventory. Therefore, the lower the out-of-stock probability, the greater the corresponding safety factor.

[Table 9]

Out of stock probability 30% 20% 15% 10% 5% Safety factor 0.53 0.85 1.04 1.29 1.65

The inventory standard setting unit 160 calculates, for each product, a demand characteristic index indicating the importance of the product based on historical data, and based on the calculated demand characteristic index and the demand prediction value of the optimal prediction algorithm obtained from the demand forecasting unit 120, Set multiple levels of inventory standards. In this implementation, the minimum inventory, recommended inventory, and maximum inventory are set.

The following describes an example in which the inventory standard setting unit 160 sets the inventory standard for the product with the product number "PT001" as the target product. The specific setting process includes the following steps (a), (b), and (c). ), (d).

(a) Find the demand characteristic index

In this embodiment, the demand characteristic indicators include demand probability and consumption ABC classification. Demand probability represents the probability of generating demand for a certain commodity within a unit period. The higher the probability, the more often the commodity is used. The consumption ABC classification is based on the ABC analysis method to classify the consumption of a certain commodity in a given period in the past, and obtain a classification result indicating the usage of the commodity. The importance of products can be distinguished based on demand probability and consumption ABC classification, and appropriate inventory standards can be set based on the importance of the products.

The methods for obtaining demand probability and consumption ABC classification are explained below.

·Demand probability

Based on the actual consumption of the commodity, the demand probability is obtained in units of a given time, such as days, weeks, or months. For example, as shown in Table 10, for the target product with product number "PT001", in monthly units, assuming that the current time point is August, the consumption data for the past year to July is obtained, and calculated on a monthly basis. Get the total consumption value of this product. Next, as shown in Table 11, for months where the total consumption value is greater than 0, fill in the value "1" indicating consumption, and for months where the total consumption value is equal to 0, fill in the value "0" indicating no consumption. Then, divide the number of months in a year with consumption by the number of months in a year to find the demand probability. Here, the number of consumption months in a year is 6, so the demand probability is 6/12 = 0.5, which means there is a 0.5 probability that there will be demand for this product in the next year. The higher the probability, the more frequent the product is. used.

[Table 10]

Warehouse ID Product Number August September October November December January February March April May June July LOC01 PT001 10 235 0 0 56 0 12 0 0 0 168 251

[Table 11]

Warehouse ID Product Number 8 use September October November December January February March April May June July LOC01 PT001 1 1 0 0 1 0 1 0 0 0 1 1

·ABC classification of consumption

FIG. 5 is a diagram showing classification based on the ABC analysis method in this embodiment. The abscissa represents the product number, the left ordinate represents the consumption of the product, and the right ordinate represents the cumulative percentage.

First, calculate the total value of the consumption of each product in the past year, and sort each product in order from the largest to the smallest total value, as shown in the multiple rectangles arranged in Figure 5. Next, the consumption of each commodity is calculated as a percentage of the overall consumption, and the cumulative percentage is calculated. The curve in Figure 5 represents the numerical change of the cumulative percentage. Products with a cumulative percentage of 0% to 80% are classified as Category A, products with a cumulative percentage of 80% to 90% are classified as Category B, and products with a cumulative percentage of 90% to 100% are classified as Category B. The goods are classified into Category C. Here, the cumulative percentage range and the number of categories can be adjusted according to the actual situation.

In Figure 5, products with product numbers 001 to 006 are category A, products with product numbers 007 to 009 are category B, and products with product numbers 010 to 015 are category C. Category A products have the highest importance, Category B products have medium importance, and Category C products have the lowest importance.

The importance of the product can be distinguished based on the numerical value of the demand probability and the classification result of the consumption ABC classification. For example, a product with a demand probability of 0.8 or more and a classification result of Class A is a product that is frequently used and used in large quantities. Such products are set as important products, and other products are set as unimportant products. Here, the threshold value of the demand probability used in the classification and the category of the classification result are not limited to these and can be changed as necessary.

(b) Calculate the minimum inventory quantity

The minimum inventory quantity indicates the minimum inventory holding quantity, and the inventory standard setting unit 160 calculates the minimum inventory quantity as follows, for example.

For important commodities with a demand probability of 0.8 or above and a classification result of Class A:

Minimum inventory = MAX (usage during the order period + safety stock corresponding to 30% out-of-stock probability, maximum consumption in a single month in the last 6 months);

For non-essential items:

Minimum inventory = MAX (usage during the order period + safety stock corresponding to 30% out-of-stock probability, average consumption in the last 6 months).

In the above formula, the prepared inventory is "usage during the order period + safety stock corresponding to 30% out-of-stock probability". For important products, select the prepared inventory and the maximum monthly consumption in the last 6 months. The larger of them is used as the minimum inventory. For non-important products, the larger of the prepared inventory and the average consumption in the last 6 months is selected as the minimum inventory. Since the maximum monthly consumption in the last 6 months is above the average consumption, the minimum inventory of important commodities is above the minimum inventory of non-important commodities.

The greater the out-of-stock probability, the lower the corresponding safety stock. Therefore, a larger "30%" out-of-stock probability is selected in the above formula for calculating the minimum inventory amount, but it is not limited to this and can be selected as needed. In addition, the consumption amount compared with the prepared inventory amount can also be set as needed.

(c) Calculate recommended inventory quantity

The recommended inventory amount indicates the recommended inventory holding amount, and the inventory standard setting unit 160 calculates the recommended inventory amount according to the following formula.

For important commodities with a demand probability of 0.8 or above and a classification result of Class A:

Recommended inventory = MAX (usage during the order period + safety stock corresponding to 15% out-of-stock probability, minimum inventory);

For non-essential items:

Recommended inventory = MAX (usage during the order period + safety stock corresponding to 20% out-of-stock probability, minimum inventory).

In the above formula, for important products, the inventory to be prepared is "the usage amount during the order period + the safety stock corresponding to the 15% out-of-stock probability", and the larger of the inventory to be prepared and the minimum inventory is selected as the recommendation inventory. For non-important products, the prepared inventory amount is "the usage amount during the order period + the safety stock corresponding to the 20% out-of-stock probability", and the larger of the prepared inventory amount and the minimum inventory amount is selected as the recommended inventory amount. According to the above formula, it is possible to ensure that the recommended inventory is above the minimum inventory.

The smaller the out-of-stock probability, the higher the corresponding safety stock. Therefore, the recommended inventory of important products is higher than the recommended inventory of non-important products. In addition, the out-of-stock probabilities of "15%" and "20%" were selected in the above formula for calculating the recommended inventory amount. You can also choose other values, but they need to be lower than the out-of-stock probability when calculating the minimum inventory amount.

(d) Calculate the maximum inventory quantity

The maximum inventory amount indicates the highest inventory holding amount, and the inventory standard setting unit 160 calculates the maximum inventory amount according to the following formula.

For important commodities with a demand probability of 0.8 or above and a classification result of Class A:

Maximum inventory = MAX (usage during the order period + safety stock corresponding to 10% out-of-stock probability, recommended inventory);

For non-essential items:

Maximum inventory = MAX (usage during the order period + safety stock corresponding to 15% out-of-stock probability, recommended inventory).

In the above formula, for important products, the prepared inventory amount is "the usage amount during the order period + the safety stock corresponding to the 10% out-of-stock probability", and the larger of the prepared inventory amount and the recommended inventory amount is selected as the highest inventory. For non-important products, the prepared inventory amount is "the usage amount during the order period + the safety stock corresponding to the 15% out-of-stock probability", and the larger of the prepared inventory amount and the recommended inventory amount is selected as the maximum inventory amount. According to the above formula, it can be ensured that the maximum inventory level is above the recommended inventory level.

The smaller the out-of-stock probability is, the higher the corresponding safety stock is. Therefore, the maximum inventory of important commodities is higher than the maximum inventory of non-important commodities. In addition, the out-of-stock probabilities of "10%" and "15%" were selected in the above formula for calculating the maximum inventory amount. You can also choose other values, but they need to be lower than the out-of-stock probability when calculating the recommended inventory amount.

Through the above steps (a), (b), (c), and (d), three inventory standards are set for the target product with the product number "PT001": the minimum inventory quantity, the recommended inventory quantity, and the maximum inventory quantity. In addition, the inventory standard is updated based on the latest demand forecast value of the target product each month, so that the inventory standard can be dynamically adjusted. For other items, follow the same steps to set and adjust inventory standards.

According to the order quantity optimization system of this embodiment, the inventory standard setting unit automatically sets a dynamic inventory standard for each product and uses the inventory standard to determine the order quantity. Therefore, it is possible to do without relying on the experience of a skilled person. The order business can be automatically executed and the order business can be realized quickly and accurately. This can improve the efficiency and accuracy of inventory management. Since inventory can be properly managed, there will be no overstocking or shortage of stock. Therefore, risks such as difficulties in capital turnover caused by inventory backlog and production shutdown caused by shortage can be prevented.

(Third Embodiment)

FIG. 6 is a block diagram showing the functional structure of the order quantity optimization system 100B according to the third embodiment of the present invention. The difference between the order quantity optimization system 100B of this embodiment and the order quantity optimization system 100A of the second embodiment is that, as shown in FIG. 6 , the order quantity optimization system 100B further includes an inventory simulation unit 170 and a replenishment plan generation unit. 180, the following explanation will focus on this difference.

<Inventory simulation unit>

The inventory simulation unit 170 is used to simulate changes in the inventory of each commodity in a given period in the future.

To this end, the inventory simulation unit 170, for each commodity, calculates the inventory change of the commodity within a given period in the future based on the current inventory, the remaining order quantity, and the demand prediction value of the optimal prediction algorithm obtained from the demand prediction unit 120. Perform simulation.

For example, an example in which the inventory quantity simulation unit 170 simulates changes in the inventory quantity of the product with the product number "PT001" as the target product will be described.

Assume that the current time point is August 2021, and the procurement cycle is 4 months. The current inventory and remaining orders are shown in Table 12 respectively. In this table, the current inventory is 300, and the next 4 months are August, The remaining orders in September, October, and November are 20, 10, 0, and 0 respectively, which means that the warehousing quantity in August is 20, the warehousing quantity in September is 10, and there is no warehousing quantity in October and November. Library.

[Table 12]

The demand prediction values of consumption in the next four months obtained from the demand prediction unit 120 by the optimal prediction algorithm are shown in Table 13. In this table, the next four months are August, September, October, and November. The demand forecast values of consumption are 165, 132, 0, and 60 respectively, which means that the predicted consumption in August is 165, the predicted consumption in September is 132, the predicted consumption in October is no consumption, and the predicted consumption in November is 60 .

[Table 13]

Warehouse ID Product Number years consumption LOC01 PT001 202108 165 LOC01 PT001 202109 132 LOC01 PT001 202110 0 LOC01 PT001 202111 60

Based on the above data, the results of the inventory changes of the simulated target commodities in the next four months are shown in Table 14. In August, the initial inventory is the current inventory at the current time point (300), the inbound quantity is the remaining order quantity for the month (20), the outbound quantity is the demand forecast value for the month (165), and the initial inventory is added Subtract the outgoing quantity from the incoming quantity (300+20-165), and the ending inventory (end-of-August inventory) is 155. In September, the initial inventory is August's ending inventory (155), the inbound quantity is the month's remaining orders (10), the outbound quantity is the month's demand forecast (132), and the incoming inventory is added to the initial inventory The inventory quantity is subtracted from the outbound quantity (155+10-132), and the ending inventory (end-of-September inventory) is 33. Calculate the ending inventory in October and November similarly, and use the ending inventory in November as the simulated inventory in the final month.

[Table 14]

Warehouse ID Product Number years initial inventory Number of warehousing Number of shipments Ending inventory LOC01 PT001 202108 300 20 165 155 LOC01 PT001 202109 155 10 132 33 LOC01 PT001 202110 33 0 0 33 LOC01 PT001 202111 33 0 60 -27

In Table 14, the ending inventory in November is negative (-27), which means that if replenishment is not carried out at the current time point, the product will be out of stock in 4 months. For such products, replenishment will be planned in the future. Priority needs to be given to ensuring replenishment when planning. Therefore, based on the results of the simulation of future inventory changes by the inventory simulation unit 170, the urgency of replenishing goods can be known, and the replenishment of goods with high urgency can be prioritized and the replenishment amount can be ensured. So as to formulate a reasonable replenishment plan.

<Replenishment plan generation unit>

The replenishment plan generation unit 180 is used to automatically formulate a replenishment plan including the replenishment quantities of all replenishment objects.

In order to formulate a more reasonable and accurate replenishment plan, the replenishment plan generation unit 180 determines, for each commodity, the replenishment priority of the commodity that currently needs to be replenished based on the simulated inventory amount of the final month obtained from the inventory simulation unit 170. The replenishment quantities of multiple stalls will eventually generate a replenishment plan including the replenishment priorities and replenishment quantities of all commodities. Since the urgency of replenishing products can be known based on the size of the simulated inventory in the last month in the inventory simulation unit 170, products with high urgency can be assigned a higher replenishment priority and set Higher replenishment volume, thereby developing a reasonable replenishment plan.

Preferably, the replenishment plan generation unit 180 also uses demand characteristic indicators to determine the replenishment priority and the replenishment amount of each gear. Since the demand characteristic index reflects the importance of the product, products with high importance can be given a higher replenishment priority and a higher replenishment quantity can be set, while products with low importance can be given a lower replenishment priority. Prioritize replenishment and set a lower replenishment quantity. Therefore, based on the inventory simulation results, demand characteristic indicators are further used to determine the replenishment priority and the replenishment volume of each gear, so that a more reasonable replenishment plan can be formulated.

Next, an example in which the replenishment plan generation unit 180 uses inventory simulation results and demand characteristic indicators to formulate a replenishment plan is described. The specific process includes the following steps (a), (b), and (c).

(a) Determine replenishment priority

The replenishment plan generation unit 180 determines the replenishment priority of the merchandise in the manner shown in Table 15. In Table 15, "S" represents the simulated inventory amount of the final month (hereinafter also referred to as simulated inventory amount), and "Other" represents the case where the demand characteristics are other than "demand probability ≥ 0.8 and consumption classification = A".

[Table 15]

For products with "S<0", the simulated inventory in the final month is negative, and the replenishment urgency is the most urgent. If replenishment is not performed at the current time point, the product will be out of stock in the final month, so regardless of the corresponding demand Regardless of the characteristics, the priority of this product is set to "High".

For products with "0≤S<minimum inventory", the simulated inventory in the final month is lower than the minimum inventory, and the replenishment urgency is urgent. If the corresponding demand characteristics are "demand probability ≥ 0.8 and consumption classification = A" , it means that the product is an important product that is frequently used and has a large usage volume, so the priority is set to "High". If the corresponding demand characteristic is "Other", it means that the product is a non-important product, so the priority is set to "High". Set to "Medium".

For products with "minimum inventory ≤ S < recommended inventory", the simulated inventory in the final month is above the minimum inventory and lower than the recommended inventory, and the replenishment urgency is not too urgent. If the corresponding demand characteristic is "demand" Probability ≥ 0.8 and consumption classification = A", it means that the product is an important product, so the priority is set to "medium". If the corresponding demand characteristic is "other", it means that the product is a non-important product, so Set priority to "Low".

For products with "recommended inventory ≤ S < maximum inventory", the simulated inventory in the final month is above the recommended inventory and lower than the maximum inventory, and the replenishment urgency is not urgent. If the corresponding demand characteristic is "demand probability ≥0.8 and consumption category=A", it means that the product is an important product, so the priority is set to "low". If the product has other demand characteristics, since it is a non-important product and the simulated inventory in the final month is high, it is not a replenishment object and is not listed in Table 15.

In addition, for products with "S ≥ maximum inventory", the simulated inventory in the final month is above the maximum inventory, so they are not subject to replenishment and are not listed in Table 15.

(b) Determine the replenishment quantity of multiple stalls

After determining the replenishment priority for the product, the replenishment plan generation unit 180 determines multiple levels of replenishment amounts for the product in the manner shown in Table 16, including the required replenishment amount, the recommended replenishment amount, and the maximum replenishment amount. Volume. In Table 16, "S" is also used to represent the simulated inventory amount in the final month, and "other" is used to represent situations other than "demand probability ≥ 0.8 and consumption classification = A".

[Table 16]

demand characteristics condition must recommend maximum Demand probability ≥ 0.8 and consumption classification = A S<0 Minimum inventory quantity-s Recommended inventory quantity-s Maximum inventory quantity-s other S<0 -S Minimum inventory quantity-s Recommended inventory quantity-s Demand probability ≥ 0.8 and consumption classification = A 0≤S＜minimum inventory quantity Minimum inventory quantity-s Recommended stock quantity-S Maximum inventory quantity-s other 0≤S＜minimum inventory quantity 0 Minimum stock quantity-S Recommended inventory quantity-s Demand probability ≥ 0.8 and consumption classification = A Minimum inventory quantity ≤S＜Recommended inventory quantity 0 Recommended stock quantity-S Maximum inventory quantity-s other Minimum inventory quantity ≤S＜Recommended inventory quantity 0 0 Recommended inventory quantity-s Demand probability ≥ 0.8 and consumption classification = A Recommended inventory quantity ≤ s < maximum inventory quantity 0 0 Maximum inventory quantity-S

For products with "S<0", the simulated inventory in the final month is negative. If replenishment is not performed at the current time point, the product will be out of stock in the final month, and the replenishment priority will be the highest. If the corresponding demand characteristics are " "Demand probability ≥ 0.8 and consumption classification = A" means that the product is an important product. Therefore, the necessary replenishment quantity is set to "minimum inventory quantity-S", that is, the minimum inventory quantity is replenished, and the recommended replenishment quantity is set to "Recommended inventory quantity-S" means to replenish the recommended inventory quantity. Set the maximum replenishment quantity to "Maximum inventory quantity-S", that is to replenish the inventory quantity to the maximum quantity. If the corresponding demand characteristic of the product is "other", it means that the product is not an important product. Therefore, the necessary replenishment quantity is set to "-S", that is, the quantity that will be out of stock in the final month is replenished, and the recommended replenishment quantity is set to It is "minimum inventory quantity - S", that is, it is replenished to the minimum inventory quantity, and the maximum replenishment quantity is set to "recommended inventory quantity - S", that is, it is replenished to the recommended inventory quantity.

For products with "0≤S<minimum inventory", the simulated inventory in the final month is lower than the minimum inventory, and the replenishment urgency is urgent. If the corresponding demand characteristics are "demand probability ≥ 0.8 and consumption classification = A" , indicating that the product is an important product, therefore, the necessary replenishment quantity is set to "minimum inventory quantity-S", that is, the minimum inventory quantity is replenished, and the recommended replenishment quantity is set to "recommended inventory quantity-S", that is, the replenishment quantity is replenished to Recommended inventory quantity, set the maximum replenishment quantity to "Maximum inventory quantity-S", that is, replenish the inventory quantity to the maximum quantity. If the corresponding demand characteristic of the product is "other", it means that the product is not an important product. Therefore, the necessary replenishment quantity is set to "0" and the recommended replenishment quantity is set to "minimum inventory quantity-S", that is, the replenishment quantity is To the minimum inventory level, set the maximum replenishment amount to "recommended inventory level -S", that is, to replenish the inventory level to the recommended inventory level.

For products with "minimum inventory ≤ S < recommended inventory", since the simulated inventory in the final month is above the minimum inventory, the replenishment quantity must be "0". If the demand characteristic corresponding to the product is "demand probability" ≥0.8 and consumption classification=A", indicating that it is an important product, set the recommended replenishment amount to "recommended inventory amount-S", that is, replenish it to the recommended inventory amount, and set the maximum replenishment amount to "maximum inventory amount-S" ”, that is, replenish the inventory to the maximum level. If the corresponding demand characteristic of the product is "other", it means that the product is a non-important product, set the recommended replenishment quantity to "0", and set the maximum replenishment quantity to "recommended inventory quantity-S", that is, replenish it to the recommended level inventory.

For products with "recommended inventory ≤ S < maximum inventory", since the simulated inventory in the final month is above the recommended inventory, both the required replenishment quantity and the recommended replenishment quantity are "0". If the corresponding demand characteristics "Demand probability ≥ 0.8 and consumption classification = A" means that the product is an important product, and the maximum replenishment amount is set to "maximum inventory - S", that is, the maximum inventory is replenished. If the product has other demand characteristics, since it is a non-important product and the simulated inventory in the final month is high, it is not a replenishment object and is not listed in Table 16.

In addition, for products with "S≥maximum inventory", since the simulated inventory in the final month is sufficient, they are not subject to replenishment and are not listed in Table 16.

For example, the minimum inventory, recommended inventory, and maximum inventory corresponding to a certain product are 100, 150, and 200 respectively, and the demand probability is 0.8, and the consumption is classified as category A. Assume that the current time is August and the procurement cycle is 4 months. If the simulated inventory of the product in November (the final month) is 50, it belongs to "0 ≤ S < minimum inventory" and "demand probability ≥ 0.8 and consumption Quantity classification=A”. According to Table 15, it can be seen that the replenishment priority of this product is "high". According to Table 16, it can be seen that the required replenishment quantity of this product is 100-50=50, the recommended replenishment quantity is 150-50=100, and the maximum replenishment quantity is 200-50=150.

If the simulated inventory of the above product in November is 120, it is a situation where "minimum inventory ≤ S < recommended inventory" and "demand probability ≥ 0.8 and consumption classification = A". According to Table 15, it can be seen that the replenishment priority of this product is "medium". According to Table 16, it can be seen that the required replenishment quantity of this product is 0, the recommended replenishment quantity is 150-120=30, and the maximum replenishment quantity is 200-120=80.

In addition, the thresholds used in classifying various situations in Table 15 and Table 16 can be appropriately changed as needed.

(c) Develop a replenishment plan

Through the above steps (a) and (b), the replenishment priority, required replenishment amount, recommended replenishment amount and maximum replenishment amount can be determined for each product. Then, the replenishment plan generation unit 180 formulates a replenishment plan including the replenishment priorities, required replenishment amounts, recommended replenishment amounts, and maximum replenishment amounts of all products targeted for replenishment. An example of a developed replenishment plan is shown in Table 17.

[Table 17]

According to the order quantity optimization system of this embodiment, future inventory changes are automatically simulated through the inventory simulation unit, and the replenishment plan generation unit automatically formulates a plan including replenishment priorities and multiple The replenishment plan of the replenishment amount of the stall can automatically execute the order business without relying on the experience of skilled workers, and the order business can be completed quickly and accurately. This can improve the efficiency and accuracy of inventory management. Since inventory can be properly managed, there will be no overstocking or shortage of stock. Therefore, risks such as difficulties in capital turnover caused by inventory backlog and production shutdown caused by shortage can be prevented.

(Fourth Embodiment)

FIG. 7 is a block diagram showing the functional structure of the order quantity optimization system 100C according to the fourth embodiment of the present invention. The difference between the order quantity optimization system 100C of this embodiment and the order quantity optimization system 100B of the third embodiment is that, as shown in FIG. 7 , the order quantity optimization system 100C also has an order plan generation unit 190. This difference will be used below. Explain the center.

Usually, the replenishment plan cannot be implemented directly as an order plan. Since there is a limit on the maximum production capacity (also referred to as production capacity in this invention) on the supplier's side, adjustments need to be made when the replenishment plan does not match the supplier's production capacity. Replenishment plan to develop final order plan.

<Order plan generation unit>

The order plan generation unit 190 is used to adjust the replenishment plan to formulate a final order plan when the replenishment plan does not match the supplier's production capacity.

There are two situations where the replenishment plan does not match the supplier's production capacity. One situation is when the total replenishment quantity exceeds the supplier's production capacity. In this case, it is necessary to reduce the replenishment quantity of some commodities so that the reduced total replenishment quantity does not exceed the supplier's production capacity. Another situation is when the total replenishment quantity is lower than the supplier's production capacity. In this case, in order to fully utilize the supplier's production capacity, it is necessary to increase the replenishment quantity of some commodities so that the increased replenishment total quantity The quantity meets the supplier’s production capacity.

After generating the replenishment plan, the order plan generation unit 190 calculates the total value of the replenishment quantities of each gear in the replenishment plan, determines whether the replenishment plan matches the supplier's production capacity, and determines if it does not match. Which situation.

Below, the two mismatch situations are explained respectively.

A. The total replenishment quantity exceeds the production capacity

Next, taking the case where the total replenishment amount exceeds the supplier's production capacity shown in Table 18 as an example, the process of the order plan generation unit 190 adjusting the replenishment plan to formulate the final order plan is explained.

In Table 18, the current time point is August 1, 2021, and the total values of the required replenishment volume, recommended replenishment volume, and maximum replenishment volume in the replenishment plan for that month are 583, 782, and 996 respectively. Assume that the supplier's production capacity in that month is 700. In this case, the total value of the recommended replenishment quantity and the maximum replenishment quantity (782, 996) exceeds the supplier's production capacity. Therefore, the order plan generation unit 190 determines that the replenishment plan does not match the supplier's production capacity, and the total replenishment amount exceeds the supplier's production capacity. The adjustment process in this case includes the following steps (a) and (b).

[Table 18]

(a) Determine the adjustment base

When the total replenishment quantity exceeds the supplier's production capacity and needs to be adjusted, it is necessary to select which replenishment quantity from multiple ranges will be used as the adjustment base.

In this implementation, if the total replenishment volume of only one stall exceeds the supplier's production capacity, the replenishment volume of that stall will be used as the adjustment base. If the total replenishment volume of more than two stalls exceeds the supply If the manufacturer's production capacity is determined, the replenishment volume of the lowest level among them will be selected as the adjustment base, that is, the level with the lowest excess amount exceeding the production capacity will be selected.

In the replenishment plan shown in Table 18, the total value of the required replenishment quantity (583) does not exceed the supplier's production capacity (700), and the total values of the recommended replenishment quantity and the maximum replenishment quantity (782, 996) exceed the supplier’s production capacity. In this case, the lowest level among the recommended replenishment quantity and the maximum replenishment quantity is the recommended replenishment quantity, and the total value exceeds the excess of the production capacity (782-700=82) and is lower than the total value of the maximum replenishment quantity. The amount is exceeded (996-700=296), so the recommended replenishment amount is selected as the adjustment base, and the adjustment is made to reduce the total value of the recommended replenishment amount (782) to the production capacity (700).

(b) Adjust replenishment quantity

In this implementation, the selected adjustment base is adjusted based on the replenishment priority and demand characteristic indicators. Below, description will be given in two cases depending on whether the total replenishment amount of the products with replenishment priorities of "medium" and "low" (hereinafter referred to as medium-low priority) is more than the excess amount.

①The total replenishment quantity of goods with medium and low replenishment priority is above the excess quantity

If the total replenishment quantity of mid- and low-priority products exceeds the excess amount, only mid- and low-priority products can be adjusted, and the total replenishment quantity of these mid- and low-priority products will be reduced. The amount is equal to the excess amount, and there is no need to adjust the replenishment amount of high-priority items. First, sort the rows in the replenishment plan according to the replenishment priority and demand characteristic indicators from low to high, and then reduce the corresponding replenishment amount starting from row 1 until the reduced total replenishment amount is supply up to the manufacturer's production capacity, or it can be reduced to a lower level. Table 19 shows an example in which the total replenishment quantity of mid- and low-priority items is adjusted when the excess quantity exceeds the excess quantity.

[Table 19]

In Table 19, the total replenishment amount of goods with medium and low replenishment priority is the sum of the recommended replenishment amounts in rows 1 to 6, which is 3+3+6+98+85+57=252. The excess amount is the replenishment amount. The difference between the total quantity of goods and the production capacity is 782-700=82, that is, the total replenishment quantity of goods with medium and low replenishment priority is greater than the excess quantity.

Moreover, in Table 19, the replenishment priority is sorted from low to high. In each row with the same replenishment priority, the demand probability and consumption ABC classification are sorted from low to high. Therefore, the closer the The upper rows correspond to lower-priority and less important products, and the corresponding recommended replenishment quantities can be reduced sequentially starting from row 1. After reducing the recommended replenishment amount in rows 1 to 3 to "0", the excess amount drops to 82-3-3-6 = 70, so then reduce the recommended replenishment amount in row 4 by 70, which is 98-70=28, the recommended replenishment quantity from line 5 onwards does not need to be adjusted, and the adjustment of the replenishment plan is completed. Taking the adjusted result as the order quantity, the last column of Table 19 shows that the total order quantity is 700, which matches the supplier's production capacity.

②The total replenishment quantity of goods with medium and low replenishment priority is less than the excess quantity

When the total replenishment quantity of mid- and low-priority commodities is less than the excess amount, adjusting only the mid- and low-priority commodities is not enough to offset the excess amount. It is necessary to further adjust the replenishment amount of high-priority commodities. Adjustment. First, sort the replenishment priority and demand characteristic indicators from low to high, and reduce the replenishment quantity of all products with medium and low replenishment priority to "0". Then, in the order of high replenishment priority, to make adjustments. When adjusting products with high replenishment priority, first ensure that the recommended replenishment amount is lower than the replenishment amount of the product. The threshold can be set appropriately as needed. Secondly, ensure that the demand probability is high and the consumption is high. The replenishment quantity of large commodities and other commodities will be adjusted according to the proportion of their respective replenishment quantities, so that the total order quantity obtained after adjustment matches the supplier's production capacity. Table 20, Table 21, and Table 22 show examples of adjustments in this case.

[Table 20]

Here, in order to illustrate that the total replenishment quantity of low- and medium-priority goods is less than the excess quantity, the supplier's production capacity is set to 500. In Table 20, the total replenishment amount of goods with medium and low replenishment priority is the sum of the recommended replenishment amounts in rows 1 to 6, which is 3+3+6+98+85+57=252. The excess amount is the replenishment amount. The difference between the total quantity of goods and the production capacity is 782-500=282, that is, the total replenishment quantity of goods with medium and low replenishment priority is less than the excess quantity.

In Table 20, the same as Table 19, the replenishment priority is sorted from low to high. In each row with the same replenishment priority, the demand probability and consumption ABC classification are sorted from low to high. The upper rows correspond to lower priority and less important items. First, reduce the replenishment quantity of all products with medium and low replenishment priority to "0", and the excess amount will be reduced to 282-252 = 30. Further adjustments need to be made for products with high replenishment priority, and then reduced by 30.

When adjusting the products with high replenishment priority in rows 7 to 13, first ensure that the recommended replenishment quantity is lower than the threshold, set the threshold to 10, and the recommended replenishment quantity of the product in row 7 (6) is lower than the threshold, therefore the recommended replenishment quantity of the item in this row is not adjusted and is directly used as the order quantity. Secondly, ensure the products with high demand probability and large consumption. The 12th to 13th products have the highest demand probability and the consumption ABC classification result is the most important category A. Therefore, the recommended replenishment quantity of the products in these rows will not be adjusted. Directly as the order quantity. For the other products in rows 8 to 11, as shown in Table 21, adjustments are made according to the proportion of the recommended replenishment quantity of the products in each row.

[Table 21]

In Table 21, for the products in rows 8 to 11, first, calculate the proportion of each row relative to the total recommended replenishment amount in rows 8 to 11. Here, the total recommended replenishment amount in rows 8 to 11 is 27 +25+35+156=243, so for example, the ratio of the goods in row 8 is 27/243≈0.11. Then, multiply the ratio of each row by the current quantity that needs to be reduced (30) to get the adjustment amount of each row, for example, the ratio of row 8 The adjustment amount of the 8 rows of products is 0.11×30≈3. Therefore, the recommended replenishment quantity of the row of products minus 3 is taken as the order quantity, which is 27-3=24. The same calculation is performed for other rows, and we get The respective order quantities are shown in the last column of Table 21.

In the example in Table 21, since only the two decimal places are retained when calculating the proportion, the subsequent parts are discarded. As a result, the final subtracted quantity is 29 but does not reach 30, and another 1 needs to be subtracted. In this case, for example, you can decrement the maximum recommended replenishment quantity in each row by 1, you can also decrement the recommended replenishment quantity corresponding to the row with the largest part that is discarded when calculating the ratio, or you can decrement the recommended replenishment quantity of any row by 1. It is recommended to reduce the replenishment quantity by 1 and adopt appropriate methods to fine-tune it. Here, fine-tuning is performed by reducing the maximum recommended replenishment quantity by 1, that is, adjusting the recommended replenishment quantity of the product in row 11 to 137-1=136. The final adjusted order quantity is shown in the last column of Table 22, with a total order quantity of 500, matching the supplier's capacity (500 in this example).

[Table 22]

B. The total replenishment quantity is lower than the production capacity

The order plan generation unit 190 calculates the total replenishment quantity of each gear in the replenishment plan. When the total replenishment quantity of each gear is lower than the supplier's production capacity, it determines that the total replenishment quantity is If the supplier's production capacity is not met, in order to fully utilize the supplier's production capacity, it is necessary to increase the replenishment volume of some commodities.

First, the maximum replenishment quantity of each commodity in the replenishment plan is used as the adjustment base. Then, in order to select the items that need to be increased in replenishment, the rows in the replenishment plan are arranged in order from the largest to the smallest inventory difference. The calculation formula of inventory difference is shown in Equation (8).

Inventory difference = historical order quantity - current inventory - estimated warehousing quantity (8)

For example, for a certain product, the cumulative order quantity in the past year is 100. As the historical order quantity, the current inventory is 20, and the expected warehousing quantity in this period is 10, then the inventory difference is 100-20-10=70. The inventory difference represents the amount that will be used in the future but has not yet been prepared in inventory. The larger the inventory difference, the more it can be increased during adjustment. Therefore, you can arrange the replenishment plan in order from large to small inventory difference, and then select a given number of rows starting from row 1 to increase the replenishment amount proportionally.

Below, Table 23 is used as an example to illustrate the process of the order plan generation unit 190 adjusting the replenishment plan to formulate the final order plan when the total replenishment quantity is lower than the supplier's production capacity.

[Table 23]

In the example shown in Table 23, the current time point is August 1, 2021. Assume that the total value of the maximum replenishment quantity in the replenishment plan for that month is 560, and the supplier's production capacity for that month is 700. In this case, the maximum replenishment quantity is used as the adjustment base, and the amount that needs to be adjusted is 700-560=140. That is, it is necessary to select a part of the goods in the replenishment plan so that the maximum replenishment quantity of this part of the goods is increased by 140 in total to meet the supplier's production capacity.

First, calculate the inventory difference of each row in the replenishment plan according to the above formula (8), and then arrange the inventory difference in order from large to small, and sort according to the demand probability and consumption ABC classification from high to low. , select the replenishment quantity of the goods in a given quantity of rows as the adjustment object, and adjust according to the proportion of the respective replenishment quantity.

Table 23 shows an example of adjusting the replenishment amounts of the products in the 1st to 10th rows after sorting. First, calculate the proportion of the replenishment amount of each row of products relative to the total replenishment amount of rows 1 to 10. Here, the total replenishment amount of rows 1 to 10 is 9166, so for example, the proportion of the product in row 1 is 3109/9166≈0.34. Then, multiply the proportion of each row of goods by the current quantity that needs to be increased (140) to get the adjustment amount of each row of goods. For example, the adjustment amount of the goods in row 1 is 0.34×140≈47, so the adjustment amount is 0.34×140≈47. The value after adding 47 to the replenishment quantity of the row product is used as the order quantity, which is 10+47=57. The same calculation is performed for other rows to obtain the respective order quantities, which are shown in the last column of Table 23. The replenishment quantity of goods in rows 1 to 10 has increased by a total of 140, thus satisfying the supplier's production capacity.

According to the order quantity optimization system of this embodiment, since the order plan generation unit automatically adjusts the replenishment plan to formulate an order plan that matches the supplier's production capacity, it is possible to automatically execute the order business without relying on the experience of a skilled person. , and can realize order business quickly and accurately. This can improve the efficiency and accuracy of inventory management. Since inventory can be properly managed, there will be no overstocking or shortage of stock. Therefore, risks such as difficulties in capital turnover caused by inventory backlog and production shutdown caused by shortage can be prevented.

(fifth embodiment)

FIG. 8 is a block diagram showing the functional structure of the order quantity optimization system 200 according to the fifth embodiment of the present invention. As shown in FIG. 8 , the order quantity optimization system 200 includes a data collection unit 210 , a demand forecast unit 220 , a memory 240 and an inventory standard setting unit 260 , which are connected to each other through a bus 250 .

Like the data collection unit 110 of the first embodiment, the data collection unit 210 collects historical data of business data related to products from the memory 240 .

The demand prediction unit 220 predicts the future demand of each commodity based on the historical data collected by the data collection unit 210. The specific prediction algorithm is not limited, and a known appropriate algorithm can be used.

The inventory standard setting unit 260 is used to dynamically set corresponding inventory standards for each product. Like the inventory standard setting unit 160 of the second embodiment, it first obtains a demand characteristic index indicating the importance of the product based on historical data. , and then multiple levels of inventory standards are set based on the calculated demand characteristic index and the demand prediction value regarding future demand obtained from the demand prediction unit 220 .

FIG. 9 is a flowchart showing the order quantity optimization method according to this embodiment. As shown in Figure 9, the order quantity optimization method of this embodiment includes a data collection step S210, a demand forecasting step S220, and an inventory standard setting step S230.

In step S210, historical data of business data related to the product is collected.

In step S220, the future demand of each commodity is predicted based on the historical data of the business data related to the commodity collected in step S210.

In step S230, for each product, a demand characteristic index indicating the importance of the product is obtained based on historical data. Based on the obtained demand characteristic index and the demand prediction value regarding future demand in step S220, a plurality of Level inventory standards.

According to the order quantity optimization system and the order quantity optimization method of this embodiment, dynamic inventory standards are automatically set for each commodity through the inventory standard setting unit or step, and the inventory standards are used to determine the order quantity. Therefore, it is possible to The order business is automatically executed without relying on the experience of a skilled person, and the order business can be realized quickly and accurately. This can improve the efficiency and accuracy of inventory management. Since inventory can be properly managed, there will be no overstocking or shortage of stock. Therefore, risks such as difficulties in capital turnover caused by inventory backlog and production shutdown caused by shortage can be prevented.

(Sixth Embodiment)

FIG. 10 is a block diagram showing the functional structure of the order quantity optimization system 300 according to the sixth embodiment of the present invention. As shown in FIG. 10 , the order quantity optimization system 300 includes a data collection unit 310 , a demand forecast unit 320 , a memory 340 , an inventory simulation unit 370 and a replenishment plan generation unit 380 , which are connected to each other through a bus 350 .

Like the data collection unit 110 of the first embodiment, the data collection unit 310 collects historical data of business data related to products from the memory 340 .

The demand prediction unit 320 predicts the future demand for each commodity based on the historical data collected by the data collection unit 310. The specific prediction algorithm is not limited, and a known appropriate algorithm can be used.

The inventory simulation unit 370 simulates changes in the inventory within a given period in the future for each commodity based on the current inventory, the remaining order quantity, and the demand prediction value regarding future demand obtained from the demand forecasting unit 320 .

For each commodity, the replenishment plan generation unit 380 determines the replenishment priority and multiple stalls of the commodity that currently need to be replenished based on the inventory amount of the final unit period in a given period in the future obtained from the inventory simulation unit 370 The replenishment amount is generated to generate a replenishment plan including the replenishment priority of all products as replenishment objects and the replenishment amount of multiple levels. For example, the replenishment plan generation unit 380 may set the replenishment priority in such a way that the smaller the inventory amount in the final unit period is, the higher the replenishment priority will be, and set the replenishment priority in each gear in such a way that the higher the replenishment priority is, the higher the replenishment priority will be. There are more ways to set the replenishment quantity of multiple stalls. The specific setting method is not limited, as long as it can generate the replenishment priority of the product and the replenishment quantity of multiple stalls.

FIG. 11 is a flowchart showing the order quantity optimization method according to this embodiment. As shown in Figure 11, the order quantity optimization method of this embodiment includes a data collection step S310, a demand forecast step S320, an inventory simulation step 330, and a replenishment plan generation step S340.

In step S310, historical data of business data related to the product is collected.

In step S320, the future demand of each commodity is predicted based on the historical data of the business data related to the commodity collected in step S310.

In step S330, for each commodity, the change in inventory in a given period in the future is simulated based on the current inventory, the remaining order quantity, and the demand prediction value about future demand in step S320.

In step S340, for each commodity, the replenishment priority of the commodity that currently needs to be replenished and the replenishment amounts of multiple stalls are determined based on the inventory amount of the final unit period in the future given period in step S330. Generate a replenishment plan that includes the replenishment priorities of all products as replenishment objects and the replenishment quantities of multiple levels.

According to the order quantity optimization system and the order quantity optimization method of this embodiment, future inventory changes are automatically simulated through the inventory simulation unit or step, and are automatically formulated through the replenishment plan generation unit or step using at least the inventory simulation results. The replenishment plan includes replenishment priorities and replenishment quantities of multiple levels, so order business can be automatically executed without relying on the experience of skilled personnel, and order business can be completed quickly and accurately. This can improve the efficiency and accuracy of inventory management. Since inventory can be properly managed, there will be no overstocking or shortage of stock. Therefore, risks such as difficulties in capital turnover caused by inventory backlog and production shutdown caused by shortage can be prevented.

(Seventh Embodiment)

FIG. 12 is a block diagram showing the functional structure of the order quantity optimization system 400 according to the seventh embodiment of the present invention. As shown in FIG. 12 , the order quantity optimization system 400 includes a data collection unit 410 , a demand forecast unit 420 , a memory 440 , a replenishment plan generation unit 480 and an order plan generation unit 490 , which are connected to each other through a bus 450 .

Like the data collection unit 110 of the first embodiment, the data collection unit 410 collects historical data of business data related to products from the memory 440 .

The demand prediction unit 420 predicts the future demand for each commodity based on the historical data collected by the data collection unit 410. The specific prediction algorithm is not limited, and a known appropriate algorithm can be used.

Replenishment plan generation unit 480 generates a replenishment plan including the replenishment amounts of all commodities targeted for replenishment, for each commodity, based on the prediction results of demand prediction unit 420 . For example, the replenishment plan generation unit 480 may directly use the prediction result of the demand forecast unit 420 as the replenishment amount, or may further consider the current inventory to set the replenishment amount, and use an appropriate method to generate the replenishment plan as needed.

The order plan generation unit 490 adjusts the replenishment quantity in the replenishment plan to match the supplier's production capacity, uses the adjusted replenishment quantity as the order quantity, and generates an order plan including the order quantity of all commodities. For example, the order plan generation unit 490 can adjust the replenishment quantities of multiple commodities with higher replenishment quantities, as long as the adjusted order quantity can match the supplier's production capacity.

FIG. 13 is a flowchart showing the order quantity optimization method according to this embodiment. As shown in Figure 13, the order quantity optimization method of this embodiment includes a data collection step S410, a demand forecasting step S420, a replenishment plan generating step S430 and an order plan generating step S440.

In step S410, historical data of business data related to the product is collected.

In step S420, the future demand of each commodity is predicted based on the historical data of the business data related to the commodity collected in step S410.

In step S430, for each commodity, a replenishment plan including the replenishment amounts of all commodities targeted for replenishment is generated based on the prediction results in step S420.

In step S440, the replenishment quantity in the replenishment plan is adjusted to match the supplier's production capacity, the adjusted replenishment quantity is used as the order quantity, and an order plan including the order quantity of all commodities is generated.

According to the order volume optimization system and the order volume optimization method of this embodiment, the replenishment plan is automatically adjusted by the order plan generation unit or step to formulate an order plan that matches the supplier's production capacity. Therefore, it is possible to do this without relying on the skill of a skilled person. Execute order business empirically and automatically, and be able to realize order business quickly and accurately. This can improve the efficiency and accuracy of inventory management. Because inventory can be properly managed, there will be no overstocking or out-of-stock situations, so it can prevent risks such as difficulties in capital turnover caused by overstocking and production shutdowns caused by out-of-stocks.

The present invention has been described above with reference to the best embodiments of the present invention. However, those skilled in the art will understand that various modifications, substitutions and changes can be made to the present invention without departing from the gist of the present invention. Various technical solutions resulting from such modifications, substitutions and changes are also included in the scope of the present invention.

Industrial availability

The order quantity optimization system, order quantity optimization method and recording medium of the present invention can be widely used in various equipment that need to optimize order quantity.

Claims (22)

1. An order volume optimization system, characterized by: The data collection unit collects historical data of business data related to products; a demand forecasting unit that uses multiple forecasting algorithms for each commodity to predict future demand based on the historical data; and A forecast algorithm selection unit, for each commodity, calculates each of the multiple forecast algorithms using the historical demand forecast values of the multiple forecast algorithms within a given period in the past and the actual demand value within the given period. According to the prediction accuracy, the prediction algorithm with the highest prediction accuracy is selected, and the demand prediction value of the prediction algorithm with the highest prediction accuracy regarding the future demand is obtained from the demand prediction unit. 2. The order quantity optimization system according to claim 1, characterized in that, The prediction algorithm selection unit calculates a plurality of prediction accuracy evaluation indicators for each prediction algorithm, and selects a prediction algorithm with the highest prediction accuracy based on a weighted moving average of the plurality of prediction accuracy evaluation indicators. 3. The order quantity optimization system according to claim 1 or 2, characterized in that, The order volume optimization system also has: The inventory standard setting unit calculates, for each product, a demand characteristic index indicating the importance of the product based on the historical data, and based on the calculated demand characteristic index and the demand prediction value obtained from the demand forecasting unit, Set inventory standards at multiple levels. 4. The order quantity optimization system according to claim 3, characterized in that, The demand characteristic indicators include demand probability and consumption ABC classification. The demand probability represents the probability of generating demand for a certain commodity within a unit period. The consumption ABC classification represents the analysis of a certain commodity in a given period in the past based on the ABC analysis method. The result of classifying the consumption within. 5. The order quantity optimization system according to claim 4, characterized in that, The inventory standard setting unit sets the commodities whose demand probability is above a given value and whose consumption ABC is classified into a given category or above as important commodities, and sets other commodities as non-important commodities, according to whether the commodities are important Commodities are used to set the inventory standards of the multiple levels respectively. 6. The order quantity optimization system according to claim 3, characterized in that, The order volume optimization system also has: The inventory simulation unit simulates, for each commodity, changes in inventory within a given period in the future based on the current inventory, remaining orders, and the demand forecast value obtained from the demand forecast unit. 7. The order quantity optimization system according to claim 6, characterized in that, The order volume optimization system also has: The replenishment plan generation unit determines, for each commodity, the replenishment priority and quantity of the commodity that currently need to be replenished based on the inventory of the final unit period in the future given period obtained from the inventory simulation unit. The replenishment amount of each stall generates a replenishment plan including the replenishment priority of all products as replenishment objects and the replenishment amounts of multiple stalls. 8. The order quantity optimization system according to claim 7, characterized in that, The replenishment plan generation unit determines the replenishment priority and the replenishment of each gear in such a way that the lower the inventory amount in the final unit period, the higher the replenishment priority and the higher the replenishment amount of each gear. quantity. 9. The order quantity optimization system according to claim 7 or 8, characterized in that, The replenishment plan generation unit also uses the demand characteristic index to determine replenishment in such a way that the higher the importance represented by the demand characteristic index, the higher the replenishment priority and the higher the replenishment amount of each gear. Priority and replenishment quantity of each gear. 10. The order quantity optimization system according to claim 7, characterized in that, The order volume optimization system also has: The order plan generation unit adjusts the replenishment quantity in the replenishment plan so as to match the supplier's production capacity, uses the adjusted replenishment quantity as the order quantity, and generates an order plan including the order quantity of all commodities. . 11. The order quantity optimization system according to claim 10, characterized in that, The order plan generation unit calculates the total value of the replenishment quantity of each gear in the replenishment plan. If there is a total value that exceeds the production capacity, based on the replenishment priority and the demand characteristic indicator, Adjust the replenishment quantity of the lowest tier among the tiers whose total value exceeds the stated production capacity. 12. The order quantity optimization system according to claim 11, characterized in that, When adjusting the replenishment amount of the lowest gear, the order plan generation unit calculates the total replenishment amount of goods with a replenishment priority below a given level, and compares the total amount with the replenishment amount of the lowest gear. The total value of the replenishment quantity exceeds the excess amount of the production capacity, In the case where the total amount is more than the excess amount, the order plan generation unit reduces the excess amount from the replenishment amount of goods below the given level, When the total quantity is lower than the excess amount, the order plan generating unit reduces the replenishment quantity of the commodities below the given level to zero, and replenishes the items with a replenishment priority higher than the given level. Commodities at a certain level, except for specific commodities, are adjusted in a manner that reduces the replenishment volume in proportion to their respective replenishment volumes. The specific commodities include commodities whose replenishment volume is lower than a given threshold and the above-mentioned Products whose demand characteristics index is above a given index. 13. The order quantity optimization system according to claim 10, characterized in that, The order plan generating unit calculates the total value of the replenishment quantity of each gear in the replenishment plan. If there is no total value that exceeds the production capacity, the highest value is calculated based on the inventory difference and the demand characteristic indicator. The replenishment quantity of the stall is adjusted, and the inventory difference is calculated by subtracting the current inventory and the expected warehousing quantity from the historical order quantity. 14. The order quantity optimization system according to claim 13, characterized in that, When adjusting the replenishment amount of the highest grade, the order plan generation unit sorts the inventory difference from large to small, and selects the demand characteristics from the product with the largest inventory difference. A given quantity of goods whose index is above a given index will be adjusted by increasing the replenishment amount in proportion to their respective inventory differences. 15. An order volume optimization system, characterized by: The data collection unit collects historical data of business data related to products; a demand forecasting unit that predicts future demand for each commodity based on the historical data; and The inventory standard setting unit calculates, for each product, a demand characteristic index indicating the importance of the product based on the historical data, and uses the obtained demand characteristic index and the future demand obtained from the demand forecasting unit demand forecast value and set multiple levels of inventory standards. 16. An order volume optimization system, characterized by: The data collection unit collects historical data of business data related to products; The demand forecasting unit predicts the future demand for each commodity based on the historical data; An inventory simulation unit, for each commodity, simulates changes in inventory within a given period in the future based on the current inventory, remaining orders, and demand forecast values about the future demand obtained from the demand forecast unit. ;as well as The replenishment plan generation unit determines, for each commodity, the replenishment priority and quantity of the commodity that currently need to be replenished based on the inventory of the final unit period in the future given period obtained from the inventory simulation unit. The replenishment amount of each stall generates a replenishment plan including the replenishment priority of all products as replenishment objects and the replenishment amounts of multiple stalls. 17. An order volume optimization system, characterized by: The data collection unit collects historical data of business data related to products; The demand forecasting unit predicts the future demand for each commodity based on the historical data; a replenishment plan generation unit that, for each commodity, generates a replenishment plan including the replenishment amounts of all commodities that are replenishment objects based on the prediction results of the demand prediction unit; and The order plan generation unit adjusts the replenishment quantity in the replenishment plan so as to match the supplier's production capacity, uses the adjusted replenishment quantity as the order quantity, and generates an order plan including the order quantity of all commodities. . 18. An order quantity optimization method, characterized by including: The data collection step collects historical data of business data related to the product; a demand forecasting step that uses multiple forecasting algorithms for each commodity to predict future demand based on the historical data; and Forecast algorithm selection step: for each commodity, use the demand forecast history values of the multiple forecast algorithms in the past given period and the actual demand value in the given period to calculate each of the multiple forecast algorithms. According to the prediction accuracy, the prediction algorithm with the highest prediction accuracy is selected, and the demand prediction value of the prediction algorithm with the highest prediction accuracy in the demand prediction step regarding the future demand is obtained. 19. An order quantity optimization method, characterized by including: The data collection step collects historical data of business data related to the product; a demand forecasting step that predicts future demand for each commodity based on the historical data; and In the inventory standard setting step, for each product, a demand characteristic index indicating the importance of the product is obtained based on the historical data, and based on the obtained demand characteristic index and the future demand in the demand forecasting step Demand forecast values set inventory standards at multiple levels. 20. An order volume optimization method, characterized by including: The data collection step collects historical data of business data related to the product; The demand forecasting step predicts the future demand for each commodity based on the historical data; An inventory simulation step, for each commodity, simulates changes in inventory within a given period in the future based on the current inventory, remaining orders, and the demand forecast value regarding the future demand in the demand forecast step; as well as In the replenishment plan generation step, for each commodity, according to the inventory amount in the final unit period in the future given period in the inventory simulation step, determine the replenishment priority and multiple replenishment priorities of the commodity that currently need to be replenished. The replenishment amount of the stall generates a replenishment plan including the replenishment priority of all products as replenishment objects and the replenishment amount of multiple stalls. 21. An order volume optimization method, characterized by including: The data collection step collects historical data of business data related to the product; The demand forecasting step predicts the future demand for each commodity based on the historical data; a replenishment plan generation step that, for each commodity, generates a replenishment plan including the replenishment quantities of all commodities as replenishment objects based on the prediction results of the demand forecast step; and The order plan generation step is to adjust the replenishment quantity in the replenishment plan so that it matches the supplier's production capacity, use the adjusted replenishment quantity as the order quantity, and generate an order plan including the order quantity of all the commodities. . 22. A computer-readable recording medium storing a program, characterized in that: The program is used to cause the computer to execute the order quantity optimization method described in any one of claims 18 to 21.