CN120949683A

CN120949683A - A method and system for intelligent cooking combination control of multi-station microwave ovens

Info

Publication number: CN120949683A
Application number: CN202511298232.4A
Authority: CN
Inventors: 曹玉保; 温国学
Original assignee: Foshan Shunde Anbenxin Rubber Products Co ltd
Current assignee: Foshan Shunde Anbenxin Rubber Products Co ltd
Priority date: 2025-09-11
Filing date: 2025-09-11
Publication date: 2025-11-14
Anticipated expiration: 2045-09-11
Also published as: CN120949683B

Abstract

The application relates to the technical field of data processing, and discloses a multi-station microwave oven intelligent cooking combination control method and system. Collecting temperature, water content and weight data of each station to establish a multi-station comprehensive state vector, inputting the state vector into a graph neural network to establish a station interaction model, determining a parallel execution scheme by adopting an out-of-order execution strategy based on the action model, backing to a stable state when the pre-calculated execution state is inconsistent, and dynamically adjusting power and time slice allocation to generate a cooperative control instruction. The application solves the problem that each station in the multi-station microwave oven lacks cooperative control and intelligent scheduling, and improves the overall efficiency of multi-station cooking and the consistency of food cooking quality.

Description

Multi-station microwave oven intelligent cooking combination control method and system

Technical Field

The application relates to the technical field of data processing, in particular to an intelligent cooking combination control method and system for a multi-station microwave oven.

Background

The existing microwave oven control technology mainly adopts a single-station independent control mode, and single food is heated through preset time and power parameters. The conventional control method is based on a fixed cooking program, a user selects a corresponding heating mode according to the type of food, and the system performs a heating process according to preset parameters. Some advanced microwave oven products incorporate sensor technology to adjust heating parameters by sensing changes in food temperature and humidity, but these technologies are still limited to single-station independent modes of operation.

However, the prior art has significant drawbacks in multi-station microwave oven applications. Firstly, an effective coordination mechanism is lacked among the stations, and power competition and heat interference are easy to generate when a plurality of stations work simultaneously, so that the cooking effect is inconsistent. Secondly, the traditional control method adopts a linear execution strategy, and the task execution sequence cannot be optimized according to the cooking requirements of different foods and the interaction relation among stations. Again, existing systems lack the ability to predict and handle abnormal conditions, and cannot be adjusted in time when certain stations deviate, affecting overall cooking quality.

Because each station independently operates and lacks mutual perceptibility, resource allocation is unreasonable and execution efficiency is low. Further analysis finds that core technical challenges in a multi-station environment include how to build an accurate model of the inter-station interaction relationship, how to implement intelligent scheduling based on task characteristics and resource constraints, how to predict and handle abnormal states during execution, and how to dynamically adjust resource allocation policies according to real-time conditions. These problems are interrelated and deepened, and the whole-flow technical innovation from data acquisition, relational modeling, task scheduling, exception handling to resource optimization is needed to be effectively solved.

Disclosure of Invention

The application provides a multi-station microwave oven intelligent cooking combination control method and system, which are used for solving the problem that each station in the multi-station microwave oven lacks cooperative control and intelligent scheduling, and improving the overall efficiency of multi-station cooking and the consistency of food cooking quality.

In a first aspect, the present application provides a method for controlling intelligent cooking combination of a multi-station microwave oven, where the method for controlling intelligent cooking combination of the multi-station microwave oven includes:

Step S1, acquiring temperature distribution data, water content data and weight change data of each station, and establishing a multi-station comprehensive state vector after data fusion processing;

s2, inputting the multi-station comprehensive state vector into a graph neural network, identifying heat transfer relations and power interaction relations among stations, and establishing a station interaction model;

s3, analyzing an execution constraint relation among cooking tasks based on the station interaction model, and rearranging starting time of each station task by adopting an out-of-order execution strategy to determine a parallel execution scheme;

Step S4, pre-calculating an execution state of the next stage aiming at each task node in the parallel execution scheme, and returning to the previous stable state when the actual execution result does not accord with the pre-calculation result;

And S5, dynamically adjusting the power distribution proportion and the time slice distribution strategy according to the actual execution progress and the resource demand condition of each station, and generating a multi-station cooperative control instruction.

Optionally, step S1 includes:

scanning and detecting the food surface in each station through an infrared temperature sensor array to obtain a two-dimensional temperature distribution matrix;

frequency scanning is carried out on food in each station based on the dielectric property detection device, so as to obtain a loss factor sequence reflecting the change of the water content in the food;

The high-precision weight sensor is arranged at the bottom of each station for continuous weighing monitoring, so as to obtain time sequence change data of the weight of the food;

And carrying out Kalman filtering denoising treatment on the two-dimensional temperature distribution matrix, the loss factor sequence and the weight time sequence change data to obtain a multi-station comprehensive state vector.

Optionally, step S2 includes:

Constructing the multi-station comprehensive state vector into a dynamic weighted graph structure, wherein each station is used as a graph node, and the difference of physical distance and power level between stations is used as an edge weight;

Performing graph convolution neural network multi-layer propagation calculation on the dynamic weighted graph structure, and obtaining node embedding vectors containing inter-station interaction information through convolution operation of an adjacent matrix and a node characteristic matrix;

Calculating heat transfer coefficients and electromagnetic interference coefficients among stations based on the node embedding vectors, and generating inter-station influence intensity weight distribution through softmax normalization processing;

and combining the heat transfer coefficient, the electromagnetic interference coefficient and the influence intensity weight distribution to construct a station interaction model matrix.

Optionally, step S3 includes:

Analyzing resource dependency relationship and time sequence constraint conditions among all cooking tasks based on the station interaction model matrix, and constructing a cooking task dependency graph through a topology sequencing algorithm;

carrying out critical path analysis on the cooking task dependency graph, and identifying an independent task set without a pre-dependency condition and an associated task chain with an execution sequence constraint;

Inputting the independent task set and the associated task chain into an out-of-order execution scheduling algorithm, and optimizing and solving a task rearrangement sequence breaking the traditional linear execution sequence through a genetic algorithm;

and generating a parallel execution scheme containing task starting time, execution duration and station allocation information according to the task rearrangement sequence and the power bearing capacity limiting condition of each station.

Optionally, the analyzing the resource dependency relationship and the time sequence constraint condition between the cooking tasks based on the station interaction model matrix, and constructing the cooking task dependency graph through a topology ordering algorithm includes:

extracting heat transfer coefficients and electromagnetic interference coefficients among stations from the station interaction model matrix, and identifying station pair combinations with mutual influence by a threshold judgment method;

Establishing a resource competition relation matrix among cooking tasks based on the station pair combination, wherein tasks in the same station pair are marked as resource conflicts, and tasks in different station pairs are marked as resources independently;

performing depth-first traversal on the resource competition relation matrix, calculating an input value and an output value of each cooking task, and generating a directed edge set reflecting the execution sequence of the tasks;

And taking each cooking task as a graph node and the directed edge set as a node connection relation, and performing topological sorting processing through a Kahn algorithm to obtain a cooking task dependency graph meeting the resource dependency constraint.

Optionally, step S4 includes:

based on the current execution state and the historical execution track of each task node in the parallel execution scheme, predicting the temperature change trend and the power demand of each station in the next time window through a long-short-term memory neural network;

Comparing the temperature change trend and the power demand with a preset execution state threshold range in a numerical value manner, and triggering an execution deviation detection mechanism when the predicted value exceeds the threshold range;

Performing state snapshot storage on a station triggering the execution deviation detection mechanism, recording power configuration parameters, temperature distribution data and task execution progress at the current moment, and constructing checkpoint state data;

And according to the check point state data, returning the execution state of the deviation station to the last historical time node of stable operation, and simultaneously updating the execution scheduling of the related task nodes in the parallel execution scheme.

Optionally, step S5 includes:

Monitoring the cooking task completion degree and the residual execution time of each station in real time, and calculating the resource utilization efficiency index of each station by combining the current power consumption condition;

Establishing a power demand priority queue based on the resource use efficiency index, setting a high-efficiency low-power station as a priority allocation object, and setting a low-efficiency high-power station as a limit allocation object;

Recalculating the power allocation quota and the time slice length of each station through a shortest job priority scheduling algorithm according to the power demand priority queue and the total power capacity limit of the system;

And packaging the recalculated power allocation quota and the time slice length into a multi-station cooperative control instruction comprising a station number, an execution instruction and a time sequence control parameter.

In a second aspect, the present application provides a multi-station microwave oven intelligent cooking combination control system, the multi-station microwave oven intelligent cooking combination control system comprising:

The fusion module is used for acquiring temperature distribution data, water content data and weight change data of each station, and establishing a multi-station comprehensive state vector after data fusion processing;

The input module is used for inputting the multi-station comprehensive state vector into a graph neural network, identifying the heat transfer relationship and the power interaction relationship among stations and establishing a station interaction model;

The starting module is used for analyzing the execution constraint relation among the cooking tasks based on the station interaction model, rearranging the starting time of each station task by adopting an out-of-order execution strategy and determining a parallel execution scheme;

The computing module is used for pre-computing the execution state of the next stage aiming at each task node in the parallel execution scheme, and returning to the previous stable state when the actual execution result is inconsistent with the pre-computing result;

the generating module is used for dynamically adjusting the power distribution proportion and the time slice distribution strategy according to the actual execution progress and the resource demand condition of each station to generate a multi-station cooperative control instruction.

In a third aspect, an intelligent cooking combination control device for a multi-station microwave oven is provided, and the intelligent cooking combination control device comprises a memory and at least one processor, wherein instructions are stored in the memory, and the at least one processor calls the instructions in the memory so that the intelligent cooking combination control device for the multi-station microwave oven can execute the intelligent cooking combination control method for the multi-station microwave oven.

In a fourth aspect, a computer readable storage medium is provided, in which instructions are stored, which when run on a computer, cause the computer to perform the above-described intelligent cooking combination control method for a multi-station microwave oven.

According to the technical scheme provided by the application, the multi-equipment layered data structure is established by collecting test parameters such as battery voltage, temperature, current and the like of each underground scraper and distributing equipment numbers, space coordinates and time stamp identifications for each data point, so that the technical problems of lack of space-time correlation and data isolation among equipment in traditional battery test data collection are effectively solved. The multi-device layered data structure is established, so that the battery test data of different devices can be subjected to association analysis under a unified space-time coordinate system, and a data base is provided for subsequent anomaly tracing and fault positioning. The multi-device hierarchical data structure is used for establishing a three-dimensional data storage model according to time dimension, space dimension and device dimension and constructing a space-time data cube supporting quick retrieval, so that the storage efficiency and the query speed of mass battery test data are remarkably improved, and compared with a traditional relational database storage mode, the space-time data cube can complete complex multi-dimensional data retrieval in millisecond-level time. When the battery state abnormality is detected, the data change track before the abnormality occurs is traced back in the space-time data cube, the propagation path of the abnormal signal among multiple devices is identified, the technical span from passive abnormal response to active abnormal prediction and tracing is realized, and the system can rapidly locate the abnormal source and analyze the influence range. Based on the propagation path, analyzing the time sequence relevance of battery state change among the devices and quantifying the direct causal influence and indirect causal influence intensity among the computing devices, the limitation that the traditional method can only identify the surface association relationship is broken through, and the accurate modeling of the deep causal relationship in the complex multi-device system is realized. According to the direct causal influence and the indirect causal influence, the collection frequency and the collection range of the battery test data of each device are adjusted, and the test data collection strategy of the cooperation of multiple devices is optimized, so that a self-adaptive data collection mechanism is formed, the system can dynamically adjust the data collection strategy according to the actual influence relation among the devices, the collection precision of key data is ensured, and the resource waste is avoided.

The space-time data cube construction algorithm enables the system to efficiently process large-scale time sequence data generated in a downhole multi-equipment environment through the design of a MapReduce parallel computing framework and a multi-level index structure, and particularly when abnormal data surge generated during equipment fault is processed, the parallel processing capacity of the algorithm ensures the instantaneity and accuracy of data processing. The abnormal propagation chain tracking algorithm adopts breadth-first search strategy and combines space-time backtracking calculation, so that the propagation path of an abnormal signal can be rapidly identified in a complex multi-equipment network. The causal relationship quantification algorithm realizes mathematical modeling of complex interrelations between devices through the combined application of Granger causal inspection and transfer entropy theory.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained based on these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic diagram of an embodiment of a control method for intelligent cooking combination of a multi-station microwave oven according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a matrix of interaction coefficients between stations according to an embodiment of the present application;

FIG. 3 is a schematic diagram of a dynamic power distribution process of a multi-station microwave oven according to an embodiment of the present application;

FIG. 4 is a schematic diagram of an embodiment of a multi-station intelligent cooking combination control system for a microwave oven in accordance with an embodiment of the present application;

fig. 5 is a schematic block diagram of a multi-station microwave oven intelligent cooking combination control apparatus according to an embodiment of the present invention.

Detailed Description

The embodiment of the application provides a multi-station microwave oven intelligent cooking combination control method and system. The terms "first," "second," "third," "fourth" and the like in the description and in the claims and in the above drawings, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments described herein may be implemented in other sequences than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.

For easy understanding, the following describes a specific flow of an embodiment of the present application, referring to fig. 1, and an embodiment of a method for controlling intelligent cooking combination of a multi-station microwave oven in an embodiment of the present application includes:

s2, inputting a multi-station comprehensive state vector into a graphic neural network, identifying heat transfer relations and power interaction relations among stations, and establishing a station interaction model;

S3, analyzing an execution constraint relation among cooking tasks based on a station interaction model, and rearranging starting time of each station task by adopting an out-of-order execution strategy to determine a parallel execution scheme;

It can be understood that the execution body of the application can be a multi-station microwave oven intelligent cooking combination control system, and can also be a terminal or a server, and the execution body is not limited in the specific description. The embodiment of the application is described by taking a server as an execution main body as an example.

Specifically, the infrared temperature sensor array scans and detects the food surface in each station, the scanning frequency is 10 times per second, a two-dimensional temperature distribution matrix is obtained, and each element in the matrix represents the temperature value of a specific position. The dielectric characteristic detection device scans the frequency of food by emitting a 2.45GHz frequency signal, detects the attenuation degree of the signal, converts the attenuation value into a loss factor value, and the loss factor directly reflects the change condition of the water content in the food. The high-precision weight sensor is arranged at the bottom of each station, continuously monitors weight change, samples at 0.1 second intervals, and records weight time sequence change data of food in the heating process. The Kalman filtering denoising processing is used for preprocessing the original sensor data, the filter is used for eliminating data noise caused by underground vibration and electromagnetic interference through a prediction step and an updating step, and the filtered data are subjected to data fusion according to a preset vector format to obtain a multi-device state vector.

And constructing a dynamic weighted graph structure by taking the multi-equipment state vector as input data, taking each station as a graph node, and calculating and determining the connection weight between the nodes according to the physical distance and the power level difference. The graph convolution neural network performs multi-layer propagation calculation on the graph structure, and each layer of convolution operation aggregates and propagates the characteristic information of adjacent nodes through mathematical operations of the adjacent matrix and the node characteristic matrix. In the forward propagation process of the network, each node collects information from neighbor nodes and updates own feature vectors after nonlinear activation function processing. And obtaining the node embedded vector containing the inter-station interaction information after multi-layer propagation. And calculating heat transfer coefficients and electromagnetic interference coefficients among stations based on the node embedding vectors, wherein the heat transfer coefficients reflect the heat energy flow intensity among the stations, and the electromagnetic interference coefficients describe the mutual influence degree of microwave power. The intensity weight distribution of the inter-station influence is generated through softmax normalization processing, and normalization ensures that the sum of all weight values is 1, so that subsequent analysis is facilitated. And combining the heat transfer coefficient, the electromagnetic interference coefficient and the influence intensity weight distribution to construct a station interaction model matrix.

And analyzing the resource dependency relationship and time sequence constraint condition among the cooking tasks based on the station interaction model matrix. The topology ordering algorithm analyzes the task dependency relationship, and identifies an independent task set without a pre-dependency condition and an associated task chain with an execution sequence constraint. Critical path analysis determines a critical task sequence that affects overall completion time by calculating the earliest and latest start times of task execution. The out-of-order execution scheduling algorithm breaks the limitation of the traditional execution according to the user input sequence, and rearranges the execution sequence according to the station resource condition and the task characteristics. The genetic algorithm is used as an optimization solving tool, the fitness function evaluation scheme is set by taking the coding task execution scheme as a chromosome, and the task rearrangement sequence is gradually optimized through the evolutionary operations such as selection, intersection, variation and the like. And after the algorithm converges, obtaining an optimal task allocation scheme, wherein the scheme considers the power bearing capacity limiting condition of each station, and generates a parallel execution scheme comprising task starting time, execution duration and station allocation information.

And establishing a prediction mechanism for each task node in the parallel execution scheme. The long-period memory neural network receives the current execution state and the historical execution track of each task node as input, and the network selectively reserves and updates the historical information through the coordination work of the forgetting gate, the input gate and the output gate so as to predict the execution state of each station in the next time window. And carrying out numerical comparison on the prediction result and a preset execution state threshold range, and triggering an execution deviation detection mechanism when the prediction value exceeds the threshold range. The state snapshot saving mechanism records power configuration parameters, temperature distribution data and task execution progress at the trigger deviation detection moment, and the data form checkpoint state data. And the rollback mechanism restores the execution state of the deviation station to the last historical time node of stable operation according to the checkpoint state data, and simultaneously updates the execution scheduling of the related task nodes in the parallel execution scheme.

And (3) establishing a real-time monitoring mechanism to track the completion degree and the residual execution time of the cooking task of each station. The resource utilization efficiency index is calculated by the ratio of the current power consumption condition to the task completion degree, and the index reflects the energy efficiency level of each station. The power demand priority queue is established according to the resource utilization efficiency index, the stations with high efficiency and low power consumption obtain the priority allocation status, and the stations with low efficiency and high power consumption are restricted by the restricted allocation. The shortest job priority scheduling algorithm recalculates the power allocation quota of each station according to the power demand priority queue and the total power capacity limit of the system, and the algorithm preferentially allocates resources for tasks with short execution time and high efficiency. The time slice length is dynamically adjusted according to the task priority, and the high-priority task obtains a longer time slice. In the dynamic adjustment process, the algorithm monitors the state change of each station in real time, and immediately recalculates the allocation scheme when the resource demand condition is found to be changed. And packaging the recalculated power allocation quota and the time slice length into control instructions, wherein the instructions comprise station numbers, execution instructions and time sequence control parameters.

In one embodiment, step S1 includes:

Frequency scanning is carried out on food in each station based on the dielectric characteristic detection device, so as to obtain a loss factor sequence reflecting the change of the water content in the food;

Specifically, in the scanning detection process of the infrared temperature sensor array, the sensors are arranged above each station in a matrix form, each sensor corresponds to a specific area of the surface of the food, and the scanning sampling mode is performed according to a preset time interval and spatial resolution. The sensor receives an infrared radiation signal emitted by the surface of the food, converts the radiation intensity into a temperature value, and the scanning process covers the whole surface of the food to obtain a two-dimensional data structure containing a plurality of temperature measurement points. Each element in the two-dimensional temperature distribution matrix corresponds to a temperature value at a specific coordinate position, and the rows and columns of the matrix represent the transverse and longitudinal position coordinates of the food surface, respectively. The acquisition of temperature data depends on the Stefan-Boltzmann law, the sensor calculates the corresponding surface temperature according to the received radiation power density, and the voltage signal output by the sensor is converted into a digital temperature value through an analog-to-digital converter.

The frequency scanning process of the dielectric characteristic detection device is based on the microwave dielectric heating principle, the device emits electromagnetic wave signals with specific frequencies to food, and the dielectric characteristics of the food are analyzed by measuring the propagation and attenuation conditions of the signals in the food. The frequency scanning means continuously changing the frequency of the transmitted signal in a preset frequency range, and recording the signal attenuation degree corresponding to each frequency point. The sequence of loss factors reflects the absorption capacity of the food for microwave signals of different frequencies, and the larger the loss factor value is, the stronger the microwave absorption capacity of the food at the frequency is. The device calculates a signal attenuation value by comparing the transmitting power and the receiving power, the attenuation value has a direct correlation with the water content of food, and the food with high water content absorbs microwave signals more strongly, so that a larger loss factor value is generated. The device records the loss factor value of each frequency point in the scanning process, and obtains a loss factor sequence according to the frequency sequence, wherein each element in the sequence corresponds to the dielectric loss characteristic under the specific frequency.

The continuous weighing monitoring of the high-precision weight sensor is realized by the strain gauge technology, and the sensor mounting position is positioned at the bottom of each station and directly bears the total weight of food and a container. And in the weighing monitoring process, the sensor continuously outputs a weight signal, the sampling interval is set to be a fixed time interval, and the weight value at the current moment is obtained through each sampling. The time sequence change data of the weight of the food records the change rule of the weight of the food along with time in the cooking process, and the weight change mainly comes from the evaporation and volatilization of water in the food. The sensor converts mechanical pressure into resistance change, the resistance change is converted into voltage signals through a Wheatstone bridge circuit, and the voltage signals are amplified by an amplifier and then input into an analog-to-digital converter to be converted into digital weight values. The continuous monitoring means that the sensor continuously works in the whole cooking process, weight data are recorded according to preset time intervals, and a data sequence containing a time stamp and a weight value is obtained.

The Kalman filtering denoising process is used as a data preprocessing link to filter random noise existing in three kinds of original data. The kalman filter algorithm estimates the true signal values through iterative processes of a prediction step and an update step based on a state space model. And the predicting step predicts the state value of the current moment according to the state estimation value of the previous moment and the dynamic system model, and the updating step combines the predicted value with the current observed value to calculate the optimal state estimation value. The algorithm maintains a covariance matrix of the state estimation during the filtering process, and the covariance matrix reflects the uncertainty degree of the estimation value. The algorithm achieves the optimal filtering effect by minimizing the variance of the estimation error, and the filtered data keeps the real signal characteristics and eliminates random noise interference. The processing process comprises the steps of independently filtering each element in the two-dimensional temperature distribution matrix, carrying out time sequence filtering on the numerical value of each frequency point in the loss factor sequence, and carrying out continuous filtering on the weight time sequence change data. And carrying out data fusion on the three types of data after filtering according to a preset vector format, and organizing different types of data into a unified data structure in a fusion process to obtain a multi-station comprehensive state vector containing temperature information, dielectric information and weight information.

In one embodiment, step S2 includes:

Constructing a multi-station comprehensive state vector into a dynamic weighted graph structure, wherein each station is used as a graph node, and the difference of physical distance and power level between stations is used as an edge weight;

Carrying out graph convolution neural network multi-layer propagation calculation on the dynamic weighted graph structure, and obtaining node embedded vectors containing inter-station interaction information through convolution operation of the adjacent matrix and the node characteristic matrix;

Calculating heat transfer coefficients and electromagnetic interference coefficients among stations based on node embedding vectors, and generating inter-station influence intensity weight distribution through softmax normalization processing;

Specifically, the dynamic weighted graph structure construction process maps each station data in the multi-station comprehensive state vector into graph nodes, and each node contains temperature distribution information, water content information and weight change information of the station. The connection relation between the nodes of the graph is represented by edge weights, and the edge weight calculation is based on two key factors of the physical distance between stations and the power level difference. The physical distance measures Euclidean distance between center points of all stations, and the closer the distance is, the stronger heat transfer and electromagnetic influence are between stations, and the larger the corresponding edge weight value is. The power level difference reflects the microwave power output difference of different stations, and the greater the power difference is, the higher the mutual interference degree between stations is, and the side weight is correspondingly increased. The dynamic characteristics are that the edge weights are adjusted along with time change, and when the station state is changed, the corresponding edge weight values are synchronously updated, so that the consistency of the graph structure and the actual physical condition is maintained.

The multi-layer propagation calculation of the graph convolution neural network is established on the basis of a dynamic weighted graph structure, and information propagation is realized through mathematical operations of an adjacent matrix and a node characteristic matrix. And recording the connection relation and weight information among all nodes in the graph by using the adjacency matrix, wherein each element in the matrix corresponds to an edge weight value between two nodes. The node characteristic matrix comprises state characteristic information of nodes of each station, rows of the matrix represent different stations, and columns represent different characteristic dimensions. And in the convolution operation process, each node collects characteristic information from adjacent nodes, the collected information is subjected to weighted summation according to the edge weights, and node characteristics are updated after the weighted summation result is processed by a nonlinear activation function. Multi-layer propagation refers to a neural network that includes multiple convolutional layers, with information propagating step by step from layer to layer, with the output of each layer serving as the input to the next layer. In the propagation process, node characteristics gradually merge information of adjacent nodes to obtain node embedded vectors containing global information. The dimension of the node embedding vector is fixed, each element in the vector represents the numerical value of the node on the specific characteristic dimension, and the embedding vector comprehensively reflects the mutual influence relationship between the station and other stations.

The node embedded vector based coefficient calculation process converts the abstract vector information into specific physical parameters. The heat transfer coefficient calculation is based on the characteristic dimension related to temperature in the node embedded vector, and vector elements are mapped into heat transfer intensity values through linear transformation. The heat transfer coefficient reflects the intensity of heat energy flow between stations, and the larger the coefficient is, the more active the heat exchange between the two stations is. The electromagnetic interference coefficient calculation is based on the characteristic dimension related to power in the node embedded vector, and the electromagnetic interference intensity value is obtained through similar linear transformation. The electromagnetic interference coefficient describes the degree of interaction of microwave power between stations, and the larger the coefficient is, the more significant the influence of power change of one station on the other station is. The softmax normalization process converts all coefficient values into a probability distribution form, and the normalization process ensures that the sum of the influence intensities of all stations on a specific station is equal to 1. The influence intensity weight distribution expresses the relative influence degree among the stations in a probability form, and each element in the weight distribution corresponds to the influence intensity proportion of a specific station pair.

The station interaction model matrix construction process combines three types of data of heat transfer coefficient, electromagnetic interference coefficient and influence intensity weight distribution into a unified mathematical model. The matrix construction adopts a block matrix form, and different areas of the matrix respectively store different types of coefficient information. The heat transfer coefficient is filled in the upper triangular area of the matrix, the electromagnetic interference coefficient is filled in the lower triangular area, and the influence intensity weight distribution is used as diagonal line elements to reflect the self influence intensity of each station. The matrix combination process comprises three links of uniform data format, standardization of numerical range and alignment of matrix dimensions. The data formats uniformly convert coefficient data from different sources into the same numerical value type, the numerical value range is standardized, the coefficients of different types are ensured to be in the same numerical value interval, and the matrix dimension alignment ensures the correct placement of each part of data in the matrix. The constructed station interaction model matrix completely describes the complex interrelationship among multiple stations.

FIG. 2 is a schematic diagram of a matrix of interaction coefficients between work stations according to an embodiment of the present application. As shown in fig. 2, the 4 x 4 matrix shows the strength of interaction between four stations (station a, station B, station C, station D). The diagonal elements of the matrix are 1.000, which represents the self-influence intensity of each station, and the off-diagonal elements reflect the interaction degree among different stations, and the larger the numerical value is, the stronger the interaction is. As can be seen from the figure, there is a strong interaction between station a and station B (coefficients 0.856 and 0.734, respectively), the coefficient of influence of station C with other stations is relatively low (in the range 0.312-0.612), and station D maintains a moderate interaction with each station. The matrix visually displays the intensity weight distribution of the inter-station influence after softmax normalization treatment through the color shade.

In one embodiment, step S3 includes:

analyzing resource dependency relationship and time sequence constraint conditions among all cooking tasks based on a station interaction model matrix, and constructing a cooking task dependency graph through a topology sequencing algorithm;

Specifically, the analysis process of the resource dependency relationship and the time sequence constraint condition is based on data information in a station interaction model matrix, and heat transfer coefficients in the matrix reflect heat interaction degrees among different cooking tasks, and electromagnetic interference coefficients describe power interaction constraint relationships among the tasks. The resource dependency relationship means that some cooking tasks can start executing after waiting for other tasks to complete or reach a specific state, and the judgment of the dependency relationship is based on the magnitude of the coefficient value in the station interaction model matrix, so that a strong dependency relationship exists between station pairs with the coefficient value exceeding a preset threshold. The timing constraints include the earliest start time, latest completion time and time interval requirements between tasks, and the constraints are determined based on the physical characteristics of the cooking process, e.g., thawing tasks must be completed before heating tasks, baking tasks cannot be performed simultaneously with high humidity tasks. The topology ordering algorithm converts these dependencies and constraints into a directed acyclic graph structure, where each node represents a cooking task and the directed edges represent the execution order requirements between the tasks. The cooking task dependency graph construction process comprises three links, namely node creation, edge connection and loop detection, wherein the node creation allocates a unique identifier for each cooking task, the edge connection establishes directed connection among related nodes according to the dependency relationship, and the loop detection ensures that no circular dependency relationship exists in the graph structure.

The critical path analysis performs depth traversal and calculation processing on the cooking task dependency graph, and the analysis process calculates the earliest starting time and the latest starting time of each task node. The earliest start time is obtained by forward computation, and the earliest possible start time of the subsequent task is computed step by step, starting from the node without the preceding task. The latest starting time is obtained through backward calculation, and the latest allowable starting time of the previous task is calculated reversely from the nodes without the subsequent task. The critical path refers to the path with the longest time consumption in all paths from the starting node to the ending node, and task nodes on the critical path are called critical tasks, and the execution time of the tasks directly affects the overall cooking completion time. Independent task sets contain task nodes that have no pre-dependency conditions and are not on critical paths, and these tasks have great flexibility in execution timing and can be adjusted according to resource availability. The associated task chain refers to a task sequence with a direct or indirect dependency relationship, the execution sequence of each task in the task chain is fixed, but the starting time of the whole task chain has an adjustment space. The identification process is realized through a graph traversal algorithm, the algorithm starts to search for depth first from each node without front dependency, all nodes on a search path are recorded, and an associated task chain is obtained.

The out-of-order execution scheduling algorithm receives an independent task set and an associated task chain as input data, and the core idea of the algorithm is to break the limitation of the traditional execution of tasks according to the input sequence of a user, and rearrange the execution sequence according to the condition of station resources and the characteristics of the tasks. The genetic algorithm is used as an optimization solving tool, the task scheduling problem is encoded into a chromosome structure of the genetic algorithm, and each chromosome represents a task execution scheme. The chromosome coding adopts the form of integer sequence, each position in the sequence corresponds to a task, and the numerical value of the position represents the station number allocated to the task. The fitness function of the genetic algorithm comprehensively considers the task completion time, the resource utilization rate and constraint violation conditions, and the chromosome with higher fitness value represents a better scheduling scheme. The algorithm generates new chromosomes through three genetic operations of selection, crossover and mutation, the selection operation selects excellent individuals to enter the next generation according to fitness values, the crossover operation exchanges gene fragments between two parent chromosomes to generate offspring, and the mutation operation randomly changes the numerical values of certain genes in the chromosomes to introduce new solutions. After multi-generation evolution, the algorithm converges to an optimal solution, a chromosome corresponding to the optimal solution is a task rearrangement sequence, the arrangement sequence of tasks in the sequence breaks through the original linear constraint, and parallelization and optimization of task execution are realized.

The parallel execution scheme generation process combines the task reordering sequence with the power bearing capacity limiting condition of each station to generate a specific execution plan. Power carrying capacity limitations include maximum power output, power regulation range, and power stability requirements for each station, which are derived from the physical characteristics and safety specifications of the microwave oven hardware. The scheme generation algorithm calculates the power distribution scheme of each station in different time periods according to the task sequence in the task rearrangement sequence and combining the power requirement and the expected execution time of each task. The task starting time calculation considers the completion time of the front-end task and the availability of the stations, ensures that the corresponding stations at the beginning of the task are in an idle state and meet the power supply requirement. The execution time length is calculated according to the task type, the food characteristics and the distribution power, the mutual influence among stations is considered in the calculation process, and the execution time is properly prolonged to compensate heat loss when adjacent stations work simultaneously. The station allocation information records the specific station number allocated to each task, the power set value and the temperature control parameter of the station during task execution.

In one embodiment, the process of analyzing the resource dependency and the time sequence constraint condition between the cooking tasks based on the station interaction model matrix in the executing step may specifically include the following steps:

extracting heat transfer coefficients and electromagnetic interference coefficients among stations from a station interaction model matrix, and identifying station pair combinations with mutual influence by a threshold judgment method;

Establishing a resource competition relation matrix among cooking tasks based on station pair combinations, wherein tasks in the same station pair are marked as resource conflicts, and tasks in different station pairs are marked as resources independently;

Performing depth-first traversal on the resource competition relation matrix, calculating the entrance value and the exit value of each cooking task, and generating a directed edge set reflecting the execution sequence of the tasks;

and taking each cooking task as a graph node and a directed edge set as a node connection relation, and performing topological sorting processing through a Kahn algorithm to obtain the cooking task dependency graph meeting the resource dependency constraint.

Specifically, a specific area for storing heat transfer coefficients and electromagnetic interference coefficients in a station interaction model matrix data extraction process positioning matrix is provided, the heat transfer coefficients are located in an upper triangular area of the matrix and reflect heat energy flow intensity between stations, the electromagnetic interference coefficients are located in a lower triangular area, and the degree of microwave power interaction between stations is described. The threshold judgment method screens out station pairs with coefficient values exceeding a preset standard by setting numerical limits, the calculation process of threshold judgment comprises direct numerical comparison of the coefficient values and the threshold, and when the heat transfer coefficient is larger than the heat threshold or the electromagnetic interference coefficient is larger than the interference threshold, the corresponding station pairs are marked as having a mutual influence relationship. The station pair combination means station pairing with mutual influence relation, and the combination data is stored in a form of a binary group and comprises numbers of two stations and influence type identifiers between the two stations, wherein the influence type identifiers distinguish two basic types of heat influence and electromagnetic influence. The extraction process is realized through a matrix traversal algorithm, the matrix elements are scanned row by row and column by the algorithm, the element positions meeting the threshold condition are recorded, and the row-column index is converted into the station number to obtain a station pair combination list.

The resource competition relation matrix establishment process converts the physical influence relation of the station layer into the resource competition relation of the task layer based on the station pair combination data. The core logic of the matrix construction is that tasks in the same station pair generate resource conflict due to sharing the same physical influence environment, and tasks in different station pairs keep a resource independent state due to influence environment independence. Resource conflict indicia refers to the existence of a resource competing relationship between cooking tasks executing within the same station pair, the root cause of the conflict being constraints of physical space and energy transfer, the cooking effects of which interfere with each other when two tasks are simultaneously executing on mutually affecting stations. Resource independent marking means that there is no direct competition for resources between tasks executing within different pairs of workstations, which can be done simultaneously without negative impact. The matrix establishment process comprises three links of matrix initialization, conflict relation filling and independent relation confirmation, wherein the matrix initialization creates a two-dimensional array with cooking tasks as row and column indexes, the conflict relation filling marks conflict states at corresponding positions according to the combined data of the station pairs, and the independent relation confirmation sets the positions without marking the conflicts as independent states. The value of each element in the matrix represents the competition strength between the corresponding task pairs, the value of the conflicting task pair is positive, and the value of the independent task pair is zero.

And performing systematic analysis on the resource competition relation matrix by using a depth-first traversal algorithm, and calculating an inbound value and an outbound value of each cooking task in the task dependency relation. Depth-first traversal refers to an algorithm strategy of starting from any task node, searching along the task dependency relationship as deeply as possible until the previous layer node is traced back when the progress cannot be continued, and searching other branches continuously. The calculation process of the income degree value counts the number of the dependency relationship pointing to the specific task, the income degree value reflects the number of the prepositioned tasks which the task needs to wait, and the task with the income degree of zero indicates that the prepositioned dependency does not exist and can immediately start to be executed. And counting the number of the dependency relationship pointing to other tasks from a specific task in the process of calculating the out-degree value, wherein the out-degree value reflects the number of subsequent tasks which can be triggered after the task is completed, and the task with zero out-degree indicates that no subsequent dependency exists and is the end point of the whole task sequence. The algorithm maintains access to the task nodes processed by the tag array record during traversal, and prevents repeated computation and infinite loops. The directed edge set generation process converts the dependency relationship among tasks into an edge structure in a graph theory, each directed edge comprises three attributes of an initial task node, a target task node and dependency strength, the dependency strength is determined according to the numerical value in the resource competition relationship matrix, and the larger the numerical value is, the stronger the dependency relationship is represented.

The Kahn algorithm topological ordering process organizes the cooking tasks and the directed edge sets into a directed acyclic graph structure, and the algorithm eliminates all edges in the graph through an iterative process to generate a task execution sequence meeting the dependency constraint. The topological sorting refers to the linear sorting of all nodes in the directed acyclic graph, and the sorting result ensures that the starting node of any directed edge in the graph is positioned before the target node in the sorting. The Kahn algorithm executing process comprises three stages of initialization, iteration processing and result output, wherein the initialization stage calculates initial degree values of all task nodes and creates a zero degree node queue, the iteration processing stage repeatedly executes the operations of node dequeuing, updating the degree of the adjacent node and adding new zero degree nodes, and the result output stage generates a topological ordering sequence. And in the execution process of the algorithm, one node is taken out of the zero-degree node queue each time and added into the sequencing result, all outgoing edges of the node are traversed, the degree value of the target node is reduced by one, and when the degree value of the target node becomes zero, the degree value of the target node is added into the zero-degree queue. After the cooking task dependency graph is constructed, all the cooking tasks are contained as graph nodes and directed edges reflecting the execution sequence, the graph structure clearly expresses complex dependency relations among the tasks, and the resource dependency constraint requirement is met.

In one embodiment, step S4 includes:

Based on the current execution state and the historical execution track of each task node in the parallel execution scheme, predicting the temperature change trend and the power demand of each station in the next time window through a long-short-period memory neural network;

Performing state snapshot storage on a station triggering an execution deviation detection mechanism, recording power configuration parameters, temperature distribution data and task execution progress at the current moment, and constructing checkpoint state data;

And according to the check point state data, the execution state of the deviation station is returned to the last historical time node of stable operation, and meanwhile, the execution time schedule of the related task nodes in the parallel execution scheme is updated.

Specifically, the speculative execution and state rollback mechanism in the intelligent cooking combination control method of the multi-station microwave oven realizes the predictive control and exception handling of the task execution state through four core data processing links. The long-period and short-period memory neural network prediction process is based on real-time state data and historical execution track data of each task node in a parallel execution scheme, and the neural network receives the current execution state as input data, wherein the input data comprises a current temperature value, a current power output value and a task completion progress percentage of each station. The historical execution track data records the change sequence of the state parameters of each station in a period of time, and the track data are arranged according to the time sequence to form a time sequence reflecting the operation mode of the station. The long-term memory neural network processes time sequence data through a special forgetting gate, an input gate and an output gate mechanism, wherein the forgetting gate decides which information is discarded from the long-term memory, the input gate decides which new information is stored in the long-term memory, and the output gate decides which information is output based on the current input and the long-term memory. The network computing process comprises three links of hidden state updating, memory unit updating and output computing, wherein the hidden state updating is combined with the hidden state of the current input and the previous moment to compute a new hidden state, the memory unit updating is used for updating long-term memory content through the coordination of a forgetting gate and an input gate, and the output computing is used for generating a prediction result based on the updated hidden state and the memory unit. The temperature change trend predicts and outputs the expected temperature value of each station in the next time window, the power demand predicts and outputs the power value required by each station to maintain the target cooking state, the predicted result is expressed in a numerical vector form, and each element in the vector corresponds to the predicted parameter of the specific station.

And performing a state threshold comparison process to carry out numerical comparison analysis on the predicted value output by the neural network and a preset safe operation range. The execution state threshold range comprises two core parts of a temperature threshold range and a power threshold range, wherein the temperature threshold range is set according to the technological requirements of different cooking tasks, and the power threshold range is determined based on the safe operation limit of microwave oven hardware. The threshold range is defined in terms of an upper limit value representing a maximum allowable value of the parameter and a lower limit value representing a minimum allowable value of the parameter. The numerical comparison process is realized through magnitude relation judgment, and when the predicted temperature change trend exceeds the temperature threshold range or the predicted power demand exceeds the power threshold range, the comparison result triggers the execution of a deviation detection mechanism. The execution deviation detection mechanism is an abnormal state identification and response mechanism, and immediately marks the corresponding station as a deviation state after the mechanism is activated, and starts the subsequent state saving and rollback flow. The judgment logic for deviation detection comprises two cases of single-parameter deviation and multi-parameter deviation, wherein the single-parameter deviation means that only the temperature or the power parameter exceeds a threshold range, and the multi-parameter deviation means that the temperature and the power parameter simultaneously exceed the respective threshold range. The detection mechanism also considers the duration of the deviation, the short-time deviation and the long-time deviation adopt different processing strategies, the short-time deviation is corrected through parameter adjustment, and the long-time deviation triggers state rollback operation.

And the state snapshot storage process carries out comprehensive state information recording and storage on the station for triggering deviation detection. The state snapshot refers to complete record of the station operation state at a specific moment, and the snapshot content comprises three core information categories of power configuration parameters, temperature distribution data and task execution progress. The power configuration parameter records the current power output set value, power regulation mode and power control strategy of the station, the parameter data is stored in the form of key value pairs, the keys represent parameter names, and the values represent specific numerical values or states of the parameters. The temperature distribution data records the temperature values of all the measuring points in the station, the data are organized in a two-dimensional array form, and the rows and columns of the array correspond to the space coordinates of the temperature measuring points respectively. Task execution progress records the percentage of completion, remaining execution time, and elapsed time of the current cooking task, the progress data being represented in a combination of a value and a time stamp. The check point state data construction process integrates three types of information according to a unified data format, and the integrated data structure comprises four fields of a time stamp, a station number, a state type and a specific numerical value. The state data storage adopts a first-in first-out caching strategy, new check point data is added to the tail of a caching queue, and old data at the front end of the caching queue is deleted when the caching space is insufficient. The buffer capacity is dynamically adjusted according to the importance of the system memory resource and the historical data, the retention time of important steady state data is longer, and the retention time of abnormal state data is relatively shorter.

The state rollback operation restores the deviation station to the historical state of the last stable operation according to the checkpoint state data. The state rollback refers to resetting the current operation parameters of the station to the parameter values of the historical stable state, and the rollback process comprises three links of parameter recovery, state verification and execution scheme updating. The parameter recovery process extracts specific parameter values of the target historical state from the checkpoint state data, resets the parameter values into corresponding station control units, and recovery operations include power output adjustment, temperature control target correction and execution of timing resets. The state verification process checks whether the state of the returned station meets the expectations, and the verification method comprises parameter value checking, state consistency checking and functional normal testing. The update process of the execution scheme adjusts task node information related to the rollback station in the parallel execution scheme, and the update content comprises the restart time of the task, the corrected execution duration and the adjusted station allocation relation. The execution scheduling update considers the influence of the rollback operation on other stations and tasks, when a certain station is in state rollback, the follow-up task depending on the station needs to delay the starting time correspondingly, and the parallel task having resource competition relation with the station needs to evaluate the execution priority again. The update algorithm determines the affected task range through the analysis of the dependency graph, recalculates the optimal execution time sequence of the tasks, and ensures the coordination and the effectiveness of the whole cooking plan.

Fig. 3 is a schematic diagram of a dynamic power distribution process of a multi-station microwave oven according to an embodiment of the present application. As shown in fig. 3, the abscissa represents cooking time, and the ordinate represents the power distribution value of each station. In the figure, the solid line represents the power distribution curve of the station a, the broken line represents the power distribution curve of the station B, the dot-dash line represents the power distribution curve of the station C, and the dotted line represents the power distribution curve of the station D. From the figure, the station A and the station B continuously distribute power in the whole cooking process, so that the characteristic of continuous cooking tasks is reflected, the power of the station C is reduced to zero after the thawing task is executed in the first 15 minutes, the resource release after the task is finished is reflected, the station D starts to distribute power after the 10 th minute, and the delay starting mechanism in the out-of-order execution strategy is reflected. The power distribution curve of each station shows a dynamic change trend, reflects a real-time adjustment process based on the resource utilization efficiency index and the power demand priority queue, and verifies that the multi-station cooperative control instruction can be dynamically and optimally distributed according to the actual execution progress and the resource demand condition of each station.

In one embodiment, step S5 includes:

Establishing a power demand priority queue based on a resource use efficiency index, setting a high-efficiency low-power-consumption station as a priority allocation object, and setting a low-efficiency high-power-consumption station as a limit allocation object;

Specifically, the real-time monitoring process continuously tracks the cooking task execution state of each station, and the monitoring data comprises three key parameters of cooking task completion, residual execution time and current power consumption. The cooking task completion is calculated by comparing the ratio of the completed cooking steps to the total cooking steps, and the completion data is expressed in percentage form to reflect the actual progress status of the task. The residual execution time is calculated based on the estimated total duration of the task minus the elapsed time, and the time data is recorded in units of minutes to dynamically reflect the time demand change of the task. The current power consumption condition is that instantaneous power consumption values of all stations are collected in real time through a power metering device, power data are expressed in watt units, and actual energy use conditions of the stations are recorded. The resource utilization efficiency index calculation process carries out correlation analysis on the task completion degree and the power consumption condition, and the calculation method is to divide the task completion degree by the corresponding power consumption accumulated value to obtain the task completion degree corresponding to the unit power consumption. The higher the efficiency index value, the larger the task amount completed by the station under the unit power consumption, and the more efficient the resource utilization. The calculation process also considers the influence of time factors, and the efficiency index is divided by the consumed time to obtain a time corrected efficiency value, so that the corrected index more accurately reflects the comprehensive resource utilization level of each station.

The power demand priority queue establishment process performs sorting and classification processing based on the resource utilization efficiency index of each station. The priority queue is a special data structure, elements in the queue are ordered according to the priority, the element with high priority is positioned at the front end of the queue, and the element with low priority is positioned at the rear end of the queue. The queue establishment process compares the efficiency indexes of all the stations to identify the stations with the highest and lowest efficiency indexes, and then sequences all the stations according to the sequence of the efficiency index values from high to low. High-efficiency low-power stations refer to stations with higher efficiency index values and relatively lower current power consumption, and the stations are set as priority allocation objects in a queue and share the priority of resource allocation. A low efficiency high power consumption workstation refers to a workstation with a low efficiency index value and a relatively high current power consumption, and such a workstation is set in a queue to limit an allocation object, and resource acquisition of the workstation is limited. The dynamic characteristics of the priority queue are shown in that the priorities of the stations can be adjusted in real time along with the change of the efficiency index, and when the efficiency index of a certain station is changed, the queue can be automatically reordered to reflect the latest priority relation. The queue management algorithm is realized by adopting a data structure of heap ordering, and the heap ordering ensures that the operations of insertion, deletion and searching of the queue have higher execution efficiency.

The shortest job priority scheduling algorithm recalculates two constraint conditions, namely, the power demand priority queue and the total power capacity limit of the system, are comprehensively considered. The shortest job priority scheduling algorithm is a classical resource scheduling strategy, and the algorithm preferentially allocates resources for the task with the shortest execution time, so that the average waiting time and the overall completion time are reduced. The core idea of the algorithm is to take the residual execution time as a measure of the length of the job, and stations with shorter residual time acquire higher resource allocation priority. The power allocation quota calculation process determines the total power capacity limit of the system and then allocates the corresponding power quota based on the location of each workstation in the priority queue and the remaining execution time. The quota allocation adopts a proportional allocation strategy, a high-priority station obtains a larger power quota, and a low-priority station obtains a smaller power quota. And determining the execution time of each station in a scheduling period according to the residual execution time and task complexity of each station in the time slice length calculation process, wherein the time slice length is in direct proportion to the priority of the station, and the station with higher priority obtains a longer time slice. And in the execution process of the algorithm, feasibility checking is needed to ensure that the sum of the power quotas of all stations does not exceed the total power capacity of the system, and when the sum of the quotas exceeds the limit, the algorithm cuts the quotas of the stations with low priority according to the priority order until the constraint condition is met.

The multi-station cooperative control instruction encapsulation process integrates the recalculated power allocation quota and the time slice length into a standardized control instruction format. The control instruction packaging is a process of converting an abstract calculation result into a specific executable instruction, and the packaged instruction comprises three core components including a station number, an execution instruction and a time sequence control parameter. The station number identifies the target execution unit of the instruction, and the number adopts a unique identifier to ensure accurate transmission and execution of the instruction. The execution instruction comprises specific power set values, operation mode selection and control strategy parameters, and instruction contents are organized in a structured data format, so that the control unit can analyze and execute the instruction conveniently. The timing control parameters specify the execution time, duration and synchronization requirements of the instructions, and the parameter settings ensure that the stations operate in accordance with coordinated timing. The instruction packaging process also comprises an instruction verification and error checking link, wherein the verification process checks the format correctness and parameter rationality of the instruction, and the error checking identifies and processes abnormal data in the instruction. The control instructions after encapsulation are sent to the control units of all stations through the communication interfaces, and after the control units receive the instructions, the control units analyze the instruction content and execute corresponding control operations. The multi-station cooperative characteristic is characterized in that control instructions of all stations are coordinated in time sequence, so that the operation of each station is ensured not to conflict or interfere with each other, and a cooperative control mechanism realizes accurate cooperation among the stations through unified clock synchronization and state monitoring.

The method for controlling intelligent cooking combination of a multi-station microwave oven in the embodiment of the present application is described above, and the system for controlling intelligent cooking combination of a multi-station microwave oven in the embodiment of the present application is described below, referring to fig. 4, an embodiment of the system for controlling intelligent cooking combination of a multi-station microwave oven in the embodiment of the present application includes:

The input module is used for inputting the multi-station comprehensive state vector into the graphic neural network, identifying the heat transfer relationship and the power interaction relationship among stations and establishing a station interaction model;

The intelligent cooking combination control system of the multi-station microwave oven in the embodiment of the invention is described in detail from the angle of modularized functional entity in fig. 4, and the intelligent cooking combination control device of the multi-station microwave oven in the embodiment of the invention is described in detail from the angle of hardware processing.

Referring to fig. 5, the embodiment of the present invention further provides a multi-station microwave oven intelligent cooking combination control device, where the multi-station microwave oven intelligent cooking combination control device may be a server, and the internal structure of the multi-station microwave oven intelligent cooking combination control device may be as shown in fig. 5. The intelligent cooking combination control equipment of the multi-station microwave oven comprises a processor, a memory, a display screen, an input device, a network interface and a database which are connected through a system bus. Wherein the computer is configured to provide computing and control capabilities. The memory of the intelligent cooking combination control equipment of the multi-station microwave oven comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the intelligent cooking combination control equipment of the multi-station microwave oven is used for storing corresponding data in the embodiment. The network interface of the intelligent cooking combination control equipment of the multi-station microwave oven is used for communicating with an external terminal through network connection. Which computer program, when being executed by a processor, carries out the above-mentioned method.

It will be understood by those skilled in the art that the structure shown in fig. 5 is only a block diagram of a portion of the structure related to the scheme of the present invention, and does not constitute a limitation of the intelligent cooking combination control apparatus of the multi-station microwave oven to which the scheme of the present invention is applied.

The invention also provides a computer readable storage medium which can be a nonvolatile computer readable storage medium, and the computer readable storage medium can also be a volatile computer readable storage medium, wherein the computer readable storage medium stores instructions which when run on a computer cause the computer to execute the steps of the intelligent cooking combination control method of the multi-station microwave oven.

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, systems and units may refer to the corresponding processes in the foregoing method embodiments, which are not repeated herein.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied essentially or partly in the form of a software product, or all or part of the technical solution, which is stored in a storage medium, and includes several instructions for causing a multi-station microwave oven intelligent cooking combination control device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. The storage medium includes a U disk, a removable hard disk, a read-only memory (ROM), a random access memory (random access memory, RAM), a magnetic disk, an optical disk, or other various media capable of storing program codes.

The foregoing embodiments are merely for illustrating the technical solution of the present invention, but not for limiting the same, and although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those skilled in the art that modifications may be made to the technical solution described in the foregoing embodiments or equivalents may be substituted for parts of the technical features thereof, and that such modifications or substitutions do not depart from the spirit and scope of the technical solution of the embodiments of the present invention in essence.

Claims

1. An intelligent cooking combination control method for a multi-station microwave oven is characterized by comprising the following steps:

2. The intelligent cooking combination control method of a multi-station microwave oven according to claim 1, wherein the step S1 includes:

3. The intelligent cooking combination control method of a multi-station microwave oven according to claim 1, wherein the step S2 includes:

4. The intelligent cooking combination control method of a multi-station microwave oven according to claim 1, wherein the step S3 includes:

5. The intelligent cooking combination control method of a multi-station microwave oven according to claim 4, wherein the analyzing the resource dependency relationship and the time sequence constraint condition between each cooking task based on the station interaction model matrix and constructing the cooking task dependency graph through the topology sequencing algorithm comprises:

6. The intelligent cooking combination control method of a multi-station microwave oven according to claim 1, wherein the step S4 includes:

7. The intelligent cooking combination control method of a multi-station microwave oven according to claim 1, wherein the step S5 includes:

8. A multi-station microwave oven intelligent cooking combination control system, characterized in that it is used for implementing a multi-station microwave oven intelligent cooking combination control method according to any one of claims 1-7, said multi-station microwave oven intelligent cooking combination control system comprising:

9. A multi-station microwave oven intelligent cooking combination control device, characterized by comprising a memory and a processor, wherein the memory stores a computer program capable of running on the processor, and the processor realizes the multi-station microwave oven intelligent cooking combination control method according to any one of claims 1 to 7 when executing the computer program.

10. A computer readable storage medium having stored thereon a computer program, characterized in that the computer program, when run by a processor, causes the processor to perform the intelligent cooking combination control method of a multi-station microwave oven according to any one of claims 1 to 7.