[go: up one dir, main page]

CN108564391B - Shared electric vehicle demand prediction method and system considering subjective and objective information - Google Patents

Shared electric vehicle demand prediction method and system considering subjective and objective information Download PDF

Info

Publication number
CN108564391B
CN108564391B CN201810021517.7A CN201810021517A CN108564391B CN 108564391 B CN108564391 B CN 108564391B CN 201810021517 A CN201810021517 A CN 201810021517A CN 108564391 B CN108564391 B CN 108564391B
Authority
CN
China
Prior art keywords
sev
demand
subjective
shared
objective
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CN201810021517.7A
Other languages
Chinese (zh)
Other versions
CN108564391A (en
Inventor
高俊杰
崔晓敏
赵鹏
陈乙庆
谢亚南
韩贤贤
王璟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dalian University of Technology
Original Assignee
Dalian University of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Dalian University of Technology filed Critical Dalian University of Technology
Priority to CN201810021517.7A priority Critical patent/CN108564391B/en
Publication of CN108564391A publication Critical patent/CN108564391A/en
Application granted granted Critical
Publication of CN108564391B publication Critical patent/CN108564391B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G06Q30/0203—Market surveys; Market polls
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G06Q30/0204—Market segmentation
    • G06Q30/0205—Market segmentation based on location or geographical consideration
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0645—Rental transactions; Leasing transactions

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Physics & Mathematics (AREA)
  • Development Economics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Economics (AREA)
  • Biophysics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Game Theory and Decision Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Physiology (AREA)
  • Genetics & Genomics (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本发明提供了一种考虑主客观信息的共享电动汽车需求预测方法及系统，SEV需求数据库包括主客观影响因素和SEV历史需求数据。通过SEV需求显著影响因素获取子系统，利用极大似然估计法提取显著影响因素。将显著影响因素和SEV历史需求数据存入SEV需求预测建模子系统，建立预测模型。将主客观预测模型存入优化SEV需求预测模型子系统，利用遗传算法优化主观logit和客观RBP的混合预测模型的权重矩阵。最后将混合预测模型存入SEV需求预测应用系统，输入主客观影响因素和SEV历史需求数据即可预测SEV需求。本发明兼顾主客观信息，准确预测SEV长期和短期需求，为SEV总量决策和车辆配送提供参考。

Figure 201810021517

The invention provides a shared electric vehicle demand forecasting method and system considering subjective and objective information. The SEV demand database includes subjective and objective influencing factors and SEV historical demand data. Through the acquisition subsystem of the significant influencing factors of SEV demand, the maximum likelihood estimation method is used to extract the significant influencing factors. The significant influencing factors and historical SEV demand data are stored in the SEV demand forecasting modeling subsystem to establish a forecasting model. The subjective and objective forecasting model is stored in the optimized SEV demand forecasting model subsystem, and the weight matrix of the mixed forecasting model of subjective logit and objective RBP is optimized by genetic algorithm. Finally, the hybrid forecast model is stored in the SEV demand forecast application system, and the SEV demand can be predicted by inputting the subjective and objective influencing factors and SEV historical demand data. The present invention takes into account subjective and objective information, accurately predicts the long-term and short-term demands of SEVs, and provides reference for SEV total amount decision-making and vehicle distribution.

Figure 201810021517

Description

Shared electric vehicle demand prediction method and system considering subjective and objective information
Technical Field
The invention relates to a demand forecasting method for a shared Electric Vehicle (SEV for short), which belongs to the field of demand forecasting of shared Electric vehicles, can reveal subjective and objective factors influencing the automobile demand of a shared site, forecast long-term or short-term automobile demand of all Electric vehicles at the shared site in advance, verify through an actual case, provide reference for total quantity determination and Vehicle distribution of the shared Electric vehicles, and effectively solve the problem of accumulated imbalance of vehicles at the shared site.
Background
With the popularization of fuel automobiles, the problems of traffic congestion, environmental pollution, difficult parking and the like in cities continuously puzzle people's traveling and daily life. The shared electric automobile can meet the traveling requirements of users, improve the utilization rate of the automobile, reduce the emission of automobile exhaust and reduce the owned quantity of private automobiles, thereby relieving traffic congestion. The subjective and objective information is comprehensively considered, and the long-term and short-term prediction accuracy of the electric vehicle demand of the shared site is improved, so that reference is provided for determining the total number of the shared vehicles of the shared electric vehicle and distributing the vehicles among the sites, the problem of accumulated imbalance of the vehicles of the shared site is effectively solved, the operator can obtain the maximum profit, the satisfaction degree of the user is maximized, and the user and the operator can achieve a win-win state.
In terms of prediction methods, the current main methods can be divided into qualitative methods and quantitative methods, wherein the qualitative prediction mainly comprises expert evaluation, questionnaire evaluation, subjective probability prediction and other methods. The quantitative prediction mainly comprises methods such as a moving average model, time series prediction, a k neighbor model, gray scale system model prediction, regression model prediction, neural network prediction and the like. Each prediction method has respective characteristics, and the prediction methods have certain results in the research of predicting the traffic flow. However, at present, research on the demand of the shared electric vehicle is still few, the demand of the shared electric vehicle has dynamics, instantaneity and uncertainty, a demand forecasting method for other goods or services cannot be well applied to the demand of the shared electric vehicle, and in addition, the demand of the shared electric vehicle is influenced by many factors, and the specific action relationship of each influencing factor is difficult to find from the demand, so that an accurate forecasting model is difficult to establish.
Disclosure of Invention
In order to effectively solve the problem of accumulation imbalance of the shared Electric vehicles, the invention provides a method and a system for predicting the demand of the shared Electric vehicles (SEV for short) by considering subjective and objective information, so as to realize long-term and short-term accurate prediction of the demand of the Electric vehicles at the shared site. In order to solve the above problems, the gist of the demand prediction for an electric vehicle according to the present invention includes:
1. and (4) building an SEV requirement database and acquiring data influencing the SEV requirement.
2. And the significant influence factor acquisition subsystem of the SEV requirement extracts the significant influence factors through a maximum likelihood estimation method.
3. The SEV demand forecasting modeling subsystem is used for establishing a forecasting model through reading subjective and objective data of significant influencing factors and historical electric vehicle demand data and respectively through a subjective logit model and an objective RBP neural network.
4. And optimizing the SEV by using a prediction model subsystem, taking the inverse of the error between the predicted value and the actual value as a fitness function, and optimizing a weight matrix of the subjective logic and objective RBP mixed prediction model by using a legacy algorithm.
5. The SEV demand forecasting application system stores the optimized electric vehicle demand forecasting model into the shared electric vehicle demand forecasting application system, inputs subjective influence factors and objective influence factors required by forecasting and electric vehicle historical demand data of a shared site, and can forecast long-term or short-term electric vehicle demands of the shared site under the condition that corresponding data are judged not to be lacked.
An SEV demand forecasting method and system considering subjective and objective information, wherein the system comprises an SEV demand database, a significant factor acquisition subsystem influencing SEV demand, an SEV demand forecasting modeling subsystem, an optimized SEV demand forecasting model subsystem and an SEV demand forecasting application system;
the SEV requirement database comprises subjective and objective factors and historical requirement data of a shared site, and the subjective factors comprise education level degree, environmental awareness, policy assistance, vehicle access convenience, charging convenience, electric vehicle brand influence and the like; objective factors comprise time periods, weeks, population density of the area, land price of the area, traffic jam conditions, public traffic coordination and the like; the public traffic coordination comprises the number of bus stations and subway stations within a service radius;
(1) the specific operation process of the SEV requirement database is as follows:
(1.1) acquiring original data related to automobile demand prediction of a shared site;
(1.2) classifying the obtained original data, and classifying the original data according to subjective factors, objective factors and historical demand of the electric automobile, wherein the subjective factors comprise education level degree, environmental awareness, policy subsidy, convenience in parking and taking the electric automobile, convenience in charging, brand influence of the electric automobile and the like; the objective factors are divided into time periods, weeks, population density of the area, land price of the area, traffic jam conditions, public traffic coordination and the like.
(1.3) representing discrete variables in the influence factors in a discrete integer form, wherein the discrete variables comprise time periods, weeks, traffic jam conditions, education level degrees, environmental awareness, policy subsidies, vehicle access convenience, charging convenience, electric automobile brand influence and the like; expressing continuous variables in the influencing factors in a continuous real number form, wherein the continuous variables comprise population density of an area, land price of the area, public traffic coordination, historical demand of the electric automobile and the like; storing the processed data in a database;
(2) the specific processing procedures for acquiring the significant factors influencing the SEV requirement are as follows:
(2.1) extracting SEV historical demand data and influence factor data;
(2.2) through the overall distribution of data, the predicted value X of the SEV requirement is a continuous real variable, and the probability distribution column of the SEV requirement is listed as
P{X＝x}＝f(x；θ1,θ2,...,θi,...,θk)
Wherein: thetaiThe i-th influencing factor required for SEV, i.e., 1,2kThe kth influencing factor of the SEV requirement; x is an actual value of the SEV demand, and X is a predicted value of the SEV demand;
(2.3) establishing a likelihood function as:
Figure BDA0001543690480000031
wherein theta isiIs the ith influencing factor of the SEV requirement; x is an actual SEV demand value;
(2.4) taking the logarithm of the likelihood function and taking its derivative, expressed as follows:
Figure BDA0001543690480000041
(2.5) solving a likelihood equation and a maximum likelihood estimation value of the influence factor, judging whether the P value of the influence factor is more than 0.05, if so, rejecting the corresponding influence factor, and performing the step (2.2) until the requirement is met; otherwise, the influence factor is kept, namely the influence factor is a significant influence factor;
(3) for a shared site i, the utility function with the SEV demand forecast value X is ViX，ViXIs a random variable which is determined by a fixed term tau in subjectively significant influencing factorsiAnd the random term psiilComposition, expressed as:
Figure BDA0001543690480000042
wherein: tau isiA fixed item of subjective significant influence factor representing SEV requirements of the shared site i; psiilThe random item of the I subjective significant influence factor representing the SEV requirement of the shared site i; e.g. of the typeilThe weight of the random item of the ith subjective significant influence factor representing the shared site i is determined by a maximum likelihood estimation method; l represents a set of random items of subjective significant influence factors of SEV requirements; subjective prediction of SEV demand based on subjective logic modelThe specific process is as follows:
(3.1) scoring the subjective significant influence factors of SEV requirements selected by the shared site in an expert scoring mode; determining the weight of the SEV demand influence factor through an evaluation matrix of the SEV demand subjective significant influence factor;
(3.2) obtaining the SEV demand probability formula of the shared site i according to the random utility maximization theory as
Figure BDA0001543690480000043
Wherein: e.g. of the typeilThe weight of the random item of the ith subjective significant influence factor representing the SEV requirement of the shared site i is determined by maximum likelihood estimation; tau isiFixed term of subjective significant influence, ψ, representing SEV requirements for shared site iilThe random item of the I subjective significant influence factor representing the SEV requirement of the shared site i; l is a set of random items of the SEV demand subjective significant influence factors; gamma is a set of shared sites;
and (3.3) obtaining the subjective logic model predicted value of the SEV requirement of the ith shared site as follows:
M1(i)＝G×ρ(i)
wherein: m1(i) The SEV demand prediction value of the subjective logic of the shared site i is obtained; g is the total number of available vehicles of the shared station i;
(4) the specific process of objective prediction of SEV demand based on objective RBP model is as follows:
(4.1) initializing an RBP neural network, and determining the total number I of nodes of the hidden layer, the center of each node, the diffusion speed and the error precision;
(4.2) inputting the SEV historical demand of the shared site and objective influence factors of the SEV demand as sample data to obtain a regression matrix P;
(4.3) selecting N groups of sample data for training, determining a weight by adopting an orthogonal least square method, and establishing a regression equation as follows:
Figure BDA0001543690480000051
wherein: d (n) is an SEV demand predicted value of the shared site of the nth iteration; i is the total number of nodes of the hidden layer; n is the number of training data; omegaiThe connection weight value of the ith hidden node and the output node is obtained; e (n) is the error between the predicted value and the actual value of the SEV requirement of the nth iteration; p is a radical ofi(n) is a regression factor of the SEV demand prediction model of the ith hidden layer in the nth iteration, a Gaussian function is selected as a basis function, and p isi(n) is represented by:
Figure BDA0001543690480000052
wherein: sigma is the standard deviation of the predicted value and the actual value of the SEV requirement; xnIs the nth training data set; t is tiIs the ith central node of the nth iteration;
(4.4) obtaining an objective predicted value M of the SEV requirement of the shared site i through the trained RBP neural network2(i)；
(5) The hybrid prediction optimization process of the SEV requirements based on the genetic algorithm mainly uses the genetic algorithm to determine the optimal weight matrix χ [ alpha, beta ] containing the SEV requirements of all shared sites],αi∈α,βiE β, i ═ 1, 2.... Γ, including initializing population, computing fitness function, individual selection, crossover and mutation operations, where α βiWeight of subjective predictor of SEV demand for shared site i, βiF is the weight of an objective predicted value of the SEV requirement of the shared site i, and is a set of the shared sites; the specific process is as follows:
(5.1) initializing population: the subjective logic predicted value and the SEV demand predicted value of the objective RBP are used as input data, and binary discrete variables are used for representing the weight alpha of the SEV demand of the shared site iiAnd betaiThe variable interval is [ c, d ]]When the discrete precision is h, the code length is log2{[d-c]/h+1}；
(5.2) calculating a fitness function: weighting alpha of SEV demand prediction of shared site i in hybrid prediction model obtained according to initial populationi、βiThe sum of the absolute values of the errors of the predicted value and the actual value is an individual fitness value f which is expressed as
Figure BDA0001543690480000061
Wherein: m1(i) SEV demand forecast for subjective locations of shared site i, M2(i) An SEV demand prediction value for objective RBP of the shared site i, S (i) an SEV actual demand value, alpha, of the shared site iiThe weight of the subjective predicted value of the SEV requirement of the shared site i; beta is aiThe weight of the objective predicted value of the SEV requirement of the shared site i; delta is a coefficient; n is the total number of individuals in the population.
(5.3) individual selection: according to the principle of roulette, individuals with smaller sums of absolute values of errors are selected.
(5.4) Cross procedure using the Single-Point Cross method, e.g., the kth chromosome akAnd the l-th chromosome alThe crossover result at point j is expressed as:
Figure BDA0001543690480000062
wherein b is [0,1 ]]The random number in the population is finally obtained, and the number of new individuals in the n individuals in the population is
Figure BDA0001543690480000063
A plurality of;
(5.5) mutation Process, selecting the jth Gene a of the ith chromosomeijAnd performing mutation, wherein the mutation result is represented as:
Figure BDA0001543690480000071
wherein: a ismaxIs gene aijMaximum value of aminIs gene aijMinimum value of r2Is [0,1 ]]Random number in interval, G is current evolution times, GmaxMaximum iteration times;
(5.6) judging whether the iteration times reach the maximum iteration times, if so, exiting the iteration process, and outputting an optimal weight matrix, otherwise, returning to the step (5.3), and continuing to evolve until the maximum iteration times are reached;
(5.7) predicting SEV requirements of the shared site i by a mixed prediction model of objective RBP and subjective logic as follows:
M(i)＝[M1(i)+M2(i)]×χ[α，β]
namely: m (i) ═ αi×M1(i)+βi×M2(i)
The SEV requirements of all shared sites predicted by the mixed prediction model of objective RBP and subjective logit are expressed as follows:
Figure BDA0001543690480000072
namely:
Figure BDA0001543690480000073
wherein: gamma is a set of shared sites; m is an SEV demand predicted value of all shared stations; m (i) is an SEV demand predicted value of the shared site i; m1(i) The subjective prediction value of the SEV requirement of the shared site i is obtained; m2(i) The objective prediction value of the SEV requirement of the shared site i is obtained; chi [ alpha, beta ]]An optimal weight matrix for SEV demand prediction of all shared sites; alpha is alphaiThe weight of the subjective predicted value of the SEV requirement of the shared site i; beta is aiThe weight of the objective predicted value of the SEV requirement of the shared site i; alpha is alphai∈α,βiE β, i ═ 1, 2.., Γ; alpha is a weight set of subjective predicted values of SEV requirements of all shared sites; β is the set of weights of the objective predicted values of SEV requirements for all shared sites.
(6) Finally, the optimized SEV demand prediction model is stored in an SEV demand prediction application system, subjective factors and objective factors required by prediction and SEV historical demand data of a shared site are input, and subjective and objective factors are input into an influence factor acquisition subsystem to acquire subjective and objective significant influence factors; and then, through a corresponding demand prediction model, under the condition of judging that no relevant data is lacked, a corresponding prediction result, namely the SEV demand can be obtained.
The invention has the beneficial effects that: the SEV demand receives the time quantum, whether be the influence of objective factors such as weekday, geographical position, population attribute, traffic conditions, also receives the influence of subjective factors such as education level degree, environmental protection consciousness, policy subsidy, access car convenience, the convenience of charging, electric automobile brand influence simultaneously. And influence factors and SEV requirements are difficult to express by using a determined analytic expression, so that the conventional single-target or multi-target optimization model is difficult to accurately predict the SEV requirements of a certain place at a certain time in the future. According to the method, subjective and objective influence factors are comprehensively considered, remarkable influence factors influencing SEV requirements are extracted through a statistical maximum likelihood estimation method, the SEV long-term or short-term requirements can be accurately predicted based on a mixed prediction model of a subjective logic model and an objective RBP model, and the accuracy of automobile requirement prediction is improved by applying artificial intelligence and heuristic algorithm technology in the field of electric automobile requirement prediction of shared sites.
Compared with a Logit multivariate prediction model and a radial basis function neural network prediction model, the hybrid optimization algorithm gives consideration to subjective influence factors and objective influence factors, and when the support popularization strength of the SEV by the national policy is increased or the education level, the environmental awareness and the like of people are raised in long-term prediction, the prediction precision can be improved by adjusting the weight of the subjective logic model; during short-term prediction, subjective factors are not changed greatly, objective influencing factors have large influence on prediction results, and prediction accuracy can be improved by adjusting the weight of an objective RBP model. And the mixed optimization prediction model has the advantages of small error, high convergence speed, less iteration times, local approximation, global optimum and the like, shows stronger self-adaption capability and robustness through the optimization of a genetic algorithm, and is suitable for predicting the SEV requirements of shared sites with large data volume and frequent calculation.
Drawings
Fig. 1 is a diagram of an SEV demand prediction organizational structure according to the present invention.
FIG. 2 is a flow chart of the SEV requirement database processing according to the present invention.
Fig. 3 is a flow chart of the present invention for extracting significant factors affecting SEV demand.
FIG. 4 is a flow chart of the SEV demand prediction model based on the subjective logic model according to the present invention.
Fig. 5 is a flow chart of SEV demand prediction based on objective RBP neural network according to the present invention.
FIG. 6 is a flow chart of the hybrid predictive model optimization of SEV demand based on genetic algorithm according to the present invention.
FIG. 7 is a flow chart of the SEV demand forecasting application system according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail by examples and drawings. The specific embodiments herein are merely illustrative of the invention and are not to be construed as limiting the invention.
An SEV demand forecasting method considering subjective and objective information comprises an SEV demand database, an SEV demand significant factor acquisition subsystem, an SEV demand forecasting modeling subsystem, an optimized SEV demand forecasting model subsystem and an SEV demand forecasting application system;
the SEV requirement database comprises subjective factors and objective factors, wherein the subjective factors comprise data such as education level degree, environmental awareness, policy assistance, vehicle access convenience, charging convenience, electric vehicle brand influence and the like; objective factors comprise time periods, week attributes, population density of the located area, price of the located area, traffic jam conditions, public traffic coordination (including the number of bus stations and subway stations in a service radius), SEV historical demand and other data.
Referring to fig. 2, the specific operation process of the SEV requirement database is as follows:
step 1, obtaining original data related to SEV demand prediction of a shared site.
Step 2, classifying the acquired original data, and classifying the original data according to subjective factors, objective factors and SEV historical demand, wherein the subjective factors are divided into data such as education level degree, environmental awareness, policy subsidy, vehicle access convenience, charging convenience, electric vehicle brand influence and the like; objective factors are divided into data such as time periods, weeks, population density of areas, land prices of the areas, traffic jam conditions, public traffic harmony (including the number of bus stations and subway stations within a service radius) and the like for separation.
Step 3, representing discrete variables (time period, week, traffic jam condition, education level degree, environmental awareness, policy subsidy, vehicle access convenience, charging convenience and electric vehicle brand influence) in the influence factors in a discrete integer form; continuous variables (population density of the located area, land price of the located area, public transportation harmony and historical demand of the electric automobile) in the influence factors are expressed in a continuous real number form.
And 4, storing the processed data into a database.
Referring to fig. 3, a specific process of extracting significant factors affecting SEV demand is as follows:
and 5, extracting historical data of the SEV requirement and data of the influencing factors.
And 6, establishing a likelihood function through the overall distribution of the data,
where SEV requirement X is a continuity variable with a probability distribution listed as
P{X＝x}＝f(x；θ1,θ2,...,θi,...,θk)
The likelihood function is established as:
Figure BDA0001543690480000101
wherein: thetaiThe i-th influencing factor required for SEV, i.e., 1,2kThe kth influencing factor of the SEV requirement; x is the actual value of the SEV demand and X is the predicted value of the SEV demand.
Step 7, taking logarithm of the likelihood function and derivation, which can be expressed as follows
Figure BDA0001543690480000102
And 8, solving the likelihood equation and the maximum likelihood estimated value of the influencing factor.
Step 9, judging whether the P value of the influence factors is more than 0.05, if so, rejecting the non-significant influence factors, and performing step 5, otherwise, performing step 10
And step 10, retaining the significant influence factors.
Referring to fig. 4, the specific process of SEV demand prediction based on the subjective logic model is as follows:
and 11, scoring the remarkable subjective influence factors selected by the shared sites in an expert scoring mode.
And 12, determining the weight of the influence factors through the evaluation matrix of the obvious subjective influence factors.
Step 13, obtaining the SEV demand probability formula of the shared site i according to the random utility maximization theory as
Figure BDA0001543690480000111
Wherein: tau isiFixed term of subjective significant influence, ψ, representing SEV requirements for shared site iil(1, 2.... 6) the first subjective significant influencer random term representing SEV demand for shared site i; l is a set of random items of the SEV demand subjective significant influence factors; gamma is the set of shared sites
Step 14, the subjective logic model predicted value of the ith shared site requirement is as follows:
M1(i)＝G×ρ(i)
wherein: m1(i) The SEV demand prediction value of the subjective logic of the shared site i is obtained; g is the total number of available vehicles of the shared station i;
referring to fig. 5, the specific process of objective prediction of SEV demand based on RBP neural network is as follows:
step 15, initializing the RBP neural network, and determining that the total number I of the nodes of the hidden layer is 10, the diffusion speed is 12, the error precision is 0.000005 and the like
And step 16, inputting a data sample of the SEV requirement of the shared site to obtain a regression matrix P.
Step 17, selecting 1571 groups of training data, learning the weight by adopting an orthogonal least square method, and establishing a regression equation as follows:
Figure BDA0001543690480000112
wherein: omegaiThe connection weight value of the ith hidden node and the output node is obtained; d (n) is an SEV demand predicted value of the shared site of the nth iteration; e (n) is the error between the predicted value and the actual value of the SEV requirement of the nth iteration; p is a radical ofi(n) is a regression factor of the SEV demand prediction model of the ith hidden layer in the nth iteration, a Gaussian function is selected as a basis function, and p isi(n) is represented by:
Figure BDA0001543690480000113
wherein, sigma is the standard deviation of the predicted value and the actual value of the SEV demand, and XnFor the nth training data set, tiIs the ith central node of the nth iteration.
Step 18, objectively predicting the SEV requirement of the shared site i to be M through the trained RBP neural network2(i)。
Referring to FIG. 6, the hybrid prediction optimization of SEV requirements based on genetic algorithm is mainly to determine the optimal weight matrix χ [ alpha, beta ] containing SEV requirements of all shared sites by using the genetic algorithm],αi∈α,βiE β, i ═ 1, 2.... Γ, including initializing population, computing fitness function, individual selection, crossover and mutation operations, where α βiWeight of subjective predictor of SEV demand for shared site i, βiF is the weight of an objective predicted value of the SEV requirement of the shared site i, and is a set of the shared sites; the specific process is as follows:
and 19, initializing the population, and taking the subjective logic predicted value and the objective RBP neural network predicted value as input data. Representing alpha by binary discrete variablesiAnd betaiOf variable quantityThe variation interval is [0,1 ]]When the discrete precision is 0.00001, the code length is log2{[1-0]0.00001+1 ≈ 16, i.e., each individual is represented by a 16-bit binary.
And 20, calculating a fitness function: weighting alpha of SEV (sequence-independent vector) requirements of shared site i of hybrid prediction model obtained according to initial populationi、βiThe sum of the absolute values of the errors of the predicted value and the actual value is an individual fitness value f which is expressed as
Figure BDA0001543690480000121
Wherein: m1(i) SEV demand forecast for subjective locations of shared site i, M2(i) An SEV demand prediction value for objective RBP of the shared site i, S (i) an SEV actual demand value, alpha, of the shared site iiThe weight of the subjective predicted value of the SEV requirement of the shared site i; beta is aiThe weight of the objective predicted value of the SEV requirement of the shared site i; delta is a coefficient; n is the total number of individuals in the population.
And 21, selecting an individual with smaller sum of absolute errors according to the roulette principle.
Step 22. crossover procedure, using single store crossover, kth chromosome akAnd the l-th chromosome alThe crossing result at point j can be expressed as:
Figure BDA0001543690480000131
wherein b is [0,1 ]]The random number in the population, and finally obtaining the new individual number among n individuals in the population as
Figure BDA0001543690480000132
And (4) respectively.
Step 23. mutation Process, selecting the jth gene a of the ith chromosomeijWhen mutation is performed, the result of mutation can be expressed as:
Figure BDA0001543690480000133
wherein, amaxIs gene aijMaximum value of aminIs gene aijMinimum value of r2Is [0,1 ]]Random number in interval, G is current evolution times, GmaxThe maximum number of iterations.
And 24, judging whether the iteration times reach the maximum iteration times.
And 25, if the maximum iteration number is met, exiting the iteration process, outputting the optimal weight matrix, otherwise returning to the step 21, and continuing to evolve until the maximum iteration number is reached.
Step 26, predicting SEV requirements of the shared site i by a mixed prediction model of objective RBP and subjective logic, and expressing the SEV requirements as follows:
M(i)＝[M1(i)+M2(i)]×χ[α，β]
namely: m (i) ═ αi×M1(i)+βi×M2(i)
The SEV requirements of all shared sites predicted by the mixed prediction model of objective RBP and subjective logit are expressed as follows:
Figure BDA0001543690480000134
namely:
Figure BDA0001543690480000135
wherein: gamma is a set of shared sites; m is an SEV demand predicted value of all shared stations; m (i) is an SEV demand predicted value of the shared site i; m1(i) The subjective predicted value of the SEV of the shared site i is obtained; m2(i) The objective predicted value of the SEV of the shared site i is obtained; chi [ alpha, beta ]]An optimal weight matrix for all shared sites; alpha is alphaiThe weight of the subjective predicted value of the SEV requirement of the shared site i; beta is aiThe weight of the objective predicted value of the SEV requirement of the shared site i; alpha is alphai∈α,βiE β, i ═ 1, 2.., Γ; alpha is a weight set of subjective predicted values of SEV requirements of all shared sites; beta is the SEV requirement of all shared sitesAnd (4) weight collection of the objective predicted value.
Referring to fig. 7, the specific process of the SEV demand forecasting application system of the shared site is as follows:
and 27, acquiring data of factors required to influence the SEV requirement and SEV historical requirement data of the shared site, wherein the influencing factors mainly comprise subjective influencing factors and objective influencing factors. Subjective influence factors include education level degree, environmental awareness, policy assistance, vehicle access convenience, charging convenience, electric vehicle brand influence and the like; the objective influence factors comprise data such as time periods, week attributes, population density of the located area, price of the located area, traffic jam conditions, public traffic harmony (including the number of bus stations and subway stations in a service radius) and the like.
And 28, inputting the subjective and objective factors of the SEV requirement and the data of the SEV requirement historical requirement into an influencing factor acquisition subsystem to acquire the remarkable subjective and objective influencing factors of the SEV requirement.
And 29, inputting the data of the significant subjective and objective influence factors of the SEV demand and the historical SEV demand into a corresponding prediction model.
And step 30, judging whether the input meets the requirement. If yes, the adjustment is continued until the requirement is met, otherwise, the step 31 is entered.
And step 31, outputting the long-term or short-term predicted value of the SEV requirement of the shared site.

Claims (1)

1. A demand forecasting method and system of a shared electric vehicle considering subjective and objective information, the shared electric vehicle is called SEV for short, and is characterized in that the system comprises an SEV demand database, an SEV demand significant influence factor acquisition subsystem, an SEV demand forecasting modeling subsystem, an optimized SEV demand forecasting model subsystem and an SEV demand forecasting application system;
the SEV requirement database comprises subjective factors and objective factors of SEV requirements, wherein the subjective factors comprise education level degree, environmental awareness, policy subsidy, vehicle access convenience, charging convenience and electric vehicle brand influence; the objective factors comprise time periods, weeks, population density of the area, land price of the area, traffic jam conditions, public traffic coordination and historical demand of the electric automobile; the public traffic coordination comprises the number of bus stations and subway stations within a service radius;
(1) the specific operation process of the SEV requirement database is as follows:
(1.1) acquiring original data related to SEV demand prediction of a shared site;
(1.2) classifying the obtained original data, and classifying the original data according to subjective factors, objective factors and historical demand of the electric automobile, wherein the subjective factors are divided into education level degree, environmental awareness, policy subsidy, convenience in parking and taking the electric automobile, convenience in charging and brand influence of the electric automobile; objective factors are divided into time periods, weeks, population density of the area, land price of the area, traffic jam conditions and public traffic coordination;
(1.3) representing discrete variables in the influence factors in a discrete integer form, wherein the discrete variables comprise time periods, weeks, traffic jam conditions, education level degrees, environmental awareness, policy subsidies, vehicle access convenience, charging convenience and electric automobile brand influence; expressing continuous variables in the influencing factors in a continuous real number form, wherein the continuous variables comprise population density of an area, land price of the area, public traffic coordination and historical demand of the electric automobile; storing the data in a database;
(2) the significant influence factors of the SEV requirement are obtained by the following specific treatment processes:
(2.1) extracting SEV historical demand data and data of SEV demand influence factors;
(2.2) through the overall distribution of data, the predicted value X of the SEV requirement is a continuous real variable, and the probability distribution column of the SEV requirement is listed as
P{X＝x}＝f(x；θ1,θ2,...,θi,...,θk)
Wherein: thetaiThe i-th influencing factor required for SEV, i.e., 1,2kThe kth influencing factor of the SEV requirement; x is an actual value of the SEV demand, and X is a predicted value of the SEV demand;
(2.3) establishing a likelihood functionComprises the following steps:
Figure FDA0002953205570000021
wherein theta isiThe kth influencing factor of the SEV requirement; x is an actual SEV demand value;
(2.4) taking the logarithm of the likelihood function and taking its derivative, expressed as follows:
Figure FDA0002953205570000022
(2.5) solving a likelihood equation and a maximum likelihood estimation value of the influence factor, judging whether the P value of the influence factor is more than 0.05, if so, rejecting the corresponding influence factor, and performing the step (2.2) until the requirement is met; otherwise, the influence factor is kept, namely the influence factor is a significant influence factor;
(3) for a shared site i, the utility function with the SEV demand forecast value X is ViX，ViXIs a random variable which is influenced by a fixed term tau in the apparent influence factoriAnd the random term psiilComposition, expressed as:
Figure FDA0002953205570000023
wherein: tau isiA subjectively significant influencer fixity representing SEV requirements for shared site i; psiilThe random item of the I subjective significant influence factor representing the SEV requirement of the shared site i; e.g. of the typeilThe weight of the random item of the ith subjective significant influence factor representing the shared site i is determined by a maximum likelihood estimation method; l represents a set of random items of the SEV requirement subjective significant influence factors; the specific process of the SEV demand subjective prediction based on the subjective logic model is as follows:
(3.1) scoring the sharing site by selecting obvious subjective influence factors of SEV requirements in an expert scoring mode; determining the weight of the SEV demand influence factor through an evaluation matrix of the subjective significant influence factor of the SEV demand;
(3.2) obtaining the SEV demand probability formula of the shared site i according to the random utility maximization theory as
Figure FDA0002953205570000031
Wherein: e.g. of the typeilThe weight of the random item of the ith subjective significant influence factor representing the shared site i is determined by a maximum likelihood estimation method; tau isiFixity of subjectively significant influencers representing SEV demand for shared site i; psiilThe random item of the I subjective significant influence factor representing the SEV requirement of the shared site i; l represents a set of random items of the SEV requirement subjective significant influence factors; gamma is a set of shared sites;
and (3.3) obtaining the subjective logic model predicted value of the SEV requirement of the ith shared site as follows:
M1(i)＝G×ρ(i)
wherein: m1(i) The SEV demand prediction value of the subjective logic of the shared site i is obtained; g is the total number of available vehicles of the shared station i;
(4) the specific process of objective prediction of SEV demand based on RBP neural network is as follows:
(4.1) initializing an RBP neural network, and determining the total number I of nodes of the hidden layer, the center of each node, the diffusion speed and the error precision;
(4.2) inputting SEV historical demand data of the shared site and objective influence factors of the SEV demand as sample data to obtain a regression matrix P;
(4.3) selecting N groups of sample data for training, learning the weight by adopting an orthogonal least square method, and establishing a regression equation as follows:
Figure FDA0002953205570000032
wherein: d (N) is the predicted value of the SEV requirement of the shared site of the nth iteration, I is the total number of nodes of the hidden layer, N is the number of training data, and omega isiIs the connection weight of the ith hidden node and the output node, e (n)) Error of predicted value and actual value of SEV demand for nth iteration, pi(n) is a regression factor of the SEV demand prediction model of the ith hidden layer in the nth iteration, a Gaussian function is selected as a basis function, and p isi(n) is represented by:
Figure FDA0002953205570000041
wherein: σ is the standard deviation of the predicted value and the actual value of the SEV demand, XnFor the nth training data set, tiIs the ith central node of the nth iteration;
(4.4) objectively predicting SEV demand objective prediction value M of shared site i through trained RBP neural network2(i)；
(5) Hybrid prediction optimization of SEV (sequence-independent vector) requirements based on genetic algorithm mainly uses the genetic algorithm to determine an optimal weight matrix χ [ alpha, beta ] containing all shared sites],αi∈α,βiE β, i ═ 1, 2.... Γ, including initializing population, computing fitness function, individual selection, crossover and mutation operations, where α βiWeight of subjective predictor of SEV demand for shared site i, βiF is the weight of an objective predicted value of the SEV requirement of the shared site i, and is a set of the shared sites; the specific process is as follows:
(5.1) initializing population: the subjective logic predicted value and the objective RBP neural network predicted value are used as input data, and binary discrete variables are used for representing the weight alpha of the SEV requirement of the shared site iiAnd betaiThe variable interval is [ c, d ]]When the discrete precision is h, the code length is log2{[d-c]/h+1}；
(5.2) calculating a fitness function: weighting alpha of SEV demand prediction of shared site i of hybrid prediction model obtained according to initial populationi、βiThe sum of the absolute values of the errors of the SEV demand predicted value and the actual value is an individual fitness value f which is expressed as
Figure FDA0002953205570000042
Wherein: m1(i) SEV demand forecast for subjective locations of shared site i, M2(i) An SEV demand prediction value for objective RBP of the shared site i, S (i) an SEV actual demand value, alpha, of the shared site iiThe weight of the subjective predicted value of the SEV requirement of the shared site i; beta is aiThe weight of the objective predicted value of the SEV requirement of the shared site i; delta is a coefficient; n is the total number of individuals in the population;
(5.3) individual selection: selecting individuals with smaller sum of absolute error values according to the roulette principle;
(5.4) Cross procedure, using the Single-Point Cross method, chromosome k akAnd the l-th chromosome alThe crossover result at point j is expressed as:
Figure FDA0002953205570000051
wherein b is [0,1 ]]The random number in the population is finally obtained, and the number of new individuals in the n individuals in the population is
Figure FDA0002953205570000052
A plurality of;
(5.5) mutation Process, selecting the jth Gene a of the ith chromosomeijAnd performing mutation, wherein the mutation result is represented as:
Figure FDA0002953205570000053
wherein: a ismaxIs gene aijMaximum value of aminIs gene aijMinimum value of r2Is [0,1 ]]Random number in interval, G is current evolution times, GmaxMaximum iteration times;
(5.6) judging whether the iteration times reach the maximum iteration times, quitting the iteration process, outputting the optimal weight matrix, and returning to (5.3) if not, continuing to evolve until the maximum iteration times are reached;
(5.7) predicting SEV requirements of the shared site i by a mixed prediction model of objective RBP and subjective logic as follows:
M(i)＝[M1(i)+M2(i)]×χ[α，β]
namely: m (i) ═ αi×M1(i)+βi×M2(i)
The SEV requirements of all shared sites predicted by the mixed prediction model of objective RBP and subjective logit are expressed as follows:
Figure FDA0002953205570000054
namely:
Figure FDA0002953205570000055
wherein: gamma is a set of shared sites; m is an SEV demand predicted value of all shared stations; m (i) is an SEV demand predicted value of the shared site i; m1(i) The subjective prediction value of the SEV requirement of the shared site i is obtained; m2(i) The objective prediction value of the SEV requirement of the shared site i is obtained; chi [ alpha, beta ]]An optimal weight matrix for SEV demand prediction of all shared sites; alpha is alphaiThe weight of the subjective predicted value of the SEV requirement of the shared site i; beta is aiThe weight of the objective predicted value of the SEV requirement of the shared site i; alpha is alphai∈α,βiE β, i ═ 1, 2.., Γ; alpha is a weight set of subjective predicted values of SEV requirements of all shared sites; beta is a weight set of objective predicted values of SEV requirements of all shared sites;
(6) finally, the optimized SEV demand prediction model is stored in an SEV demand prediction application system, subjective factors and objective factors required by the SEV demand to be predicted and SEV demand historical data of a shared site are input, the influence factors are input into an influence factor acquisition subsystem, and the obvious influence factors are acquired; and then, through a corresponding SEV demand prediction model, under the condition of judging that no relevant data is lacked, a corresponding prediction result, namely the SEV demand can be obtained.
CN201810021517.7A 2018-01-10 2018-01-10 Shared electric vehicle demand prediction method and system considering subjective and objective information Expired - Fee Related CN108564391B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810021517.7A CN108564391B (en) 2018-01-10 2018-01-10 Shared electric vehicle demand prediction method and system considering subjective and objective information

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810021517.7A CN108564391B (en) 2018-01-10 2018-01-10 Shared electric vehicle demand prediction method and system considering subjective and objective information

Publications (2)

Publication Number Publication Date
CN108564391A CN108564391A (en) 2018-09-21
CN108564391B true CN108564391B (en) 2021-05-18

Family

ID=63529766

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810021517.7A Expired - Fee Related CN108564391B (en) 2018-01-10 2018-01-10 Shared electric vehicle demand prediction method and system considering subjective and objective information

Country Status (1)

Country Link
CN (1) CN108564391B (en)

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109948898B (en) * 2019-02-02 2023-09-19 中国北方车辆研究所 A mobile robot selection method based on demand probability
CN110119838A (en) * 2019-04-17 2019-08-13 成都信息工程大学 A kind of shared bicycle demand forecast system, method and device
CN110348084B (en) * 2019-06-26 2022-08-19 西南交通大学 Subway vehicle technical index dynamic weight acquisition system and method for multi-subject customers
CN112381560B (en) * 2020-10-23 2022-10-21 东北石油大学 A kind of shared equipment product market forecasting system and method
CN112819239B (en) * 2021-02-19 2024-05-14 北京骑胜科技有限公司 Product quantity prediction method, device, electronic equipment and readable storage medium
CN113360831B (en) * 2021-06-01 2022-06-10 深圳市中晴云科技有限公司 An intelligent logistics information system based on the Internet of Things
CN114626766B (en) * 2022-05-12 2022-08-23 深圳市伟创高科电子有限公司 Shared electric vehicle scheduling method, device, equipment and medium based on big data
CN115796404A (en) * 2022-11-10 2023-03-14 国网江苏省电力有限公司营销服务中心 A power consumption forecasting method and system for agency power purchase users
CN116911911B (en) * 2023-09-12 2024-05-28 杭州慧泰数据科技有限公司 A method and system for predicting the release of public products

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6876955B1 (en) * 2001-12-28 2005-04-05 Fannie Mae Method and apparatus for predicting and reporting a real estate value based on a weighted average of predicted values
CN103115779A (en) * 2012-10-24 2013-05-22 中国电力科学研究院 Electric automobile on-line state monitoring system
CN103632205A (en) * 2013-11-05 2014-03-12 常州大学 Optimized electric-vehicle dispatching method considering wind-electricity and load uncertainty
CN106022514A (en) * 2016-05-12 2016-10-12 华南理工大学 Public electric bicycle leasing point address-selecting method based on trip chain

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6876955B1 (en) * 2001-12-28 2005-04-05 Fannie Mae Method and apparatus for predicting and reporting a real estate value based on a weighted average of predicted values
CN103115779A (en) * 2012-10-24 2013-05-22 中国电力科学研究院 Electric automobile on-line state monitoring system
CN103632205A (en) * 2013-11-05 2014-03-12 常州大学 Optimized electric-vehicle dispatching method considering wind-electricity and load uncertainty
CN106022514A (en) * 2016-05-12 2016-10-12 华南理工大学 Public electric bicycle leasing point address-selecting method based on trip chain

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
The Added Value of Accounting For Users’;Correia G H D A等;《Journal of Intelligent Transportation Systems》;20141231;全文 *
Understanding the factors affecting vehicle usage and;De Lorimier A等;《International Journal of Sustainable Transportation》;20131231;全文 *
城市公共自行车站点需求预测及调度优化方法研究;刘路美;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20170615;全文 *
基于出行链随机模拟的电动汽车充电需求分析;温剑锋等;《电网技术》;20150605;全文 *
基于电动汽车出行随机模拟的充电桩需求研究;麻秀范等;《电工技术学报》;20171027;全文 *

Also Published As

Publication number Publication date
CN108564391A (en) 2018-09-21

Similar Documents

Publication Publication Date Title
CN108564391B (en) Shared electric vehicle demand prediction method and system considering subjective and objective information
Straka et al. Predicting popularity of electric vehicle charging infrastructure in urban context
Wang Predicting tourism demand using fuzzy time series and hybrid grey theory
CN110674999A (en) Residential load prediction method based on improved clustering and long short-term memory deep learning
Pitombo et al. Comparing decision tree algorithms to estimate intercity trip distribution
Liu et al. Exploring the impact of spatiotemporal granularity on the demand prediction of dynamic ride-hailing
CN114066503A (en) Method for forecasting taxi demand based on virtual service area construction
CN117933512B (en) Get-off point recommendation method, device, computer equipment and storage medium
CN109214863A (en) A method of based on fast delivery data predicted city house demand
CN115115389A (en) A method for predicting express customer churn based on value segmentation and integrated forecasting
CN114943356A (en) Airport arrival passenger taxi short-time demand integrated prediction method
CN107316096A (en) A kind of track traffic one-ticket pass passenger amount of entering the station Forecasting Methodology
Xu et al. A taxi dispatch system based on prediction of demand and destination
CN115034429A (en) Recommendation method of city taxi pick-up point based on deep graph neural network
Guo et al. A prediction-based iterative Kuhn-Munkres approach for service vehicle reallocation in ride-hailing
CN112149976B (en) Power grid accurate investment project decision method based on artificial intelligence
Teodorović et al. A real-time road pricing system: The case of a two-link parallel network
CN116822813A (en) Task allocation method, device, equipment and storage medium in space crowdsourcing
CN118761886B (en) Passenger evacuation method, device, equipment, storage medium and computer program product
Ma et al. Research and application of second-hand housing price prediction model based on lstm
CN120087600A (en) An index system construction and evaluation method for urban green travel level
CN117455038A (en) An intelligent prediction method for short-term passenger flow in urban rail transit
Yang et al. Multi-model bus arrival time prediction using real-time online information
CN115965141A (en) Taxi demand prediction system method based on deep learning and information entropy
CN114463978A (en) Data monitoring method based on rail transit information processing terminal

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20210518

CF01 Termination of patent right due to non-payment of annual fee