CN114037480A - New-energy vehicle charging pile demand prediction and deployment optimization method for new city - Google Patents

New-energy vehicle charging pile demand prediction and deployment optimization method for new city Download PDF

Info

Publication number
CN114037480A
CN114037480A CN202111390832.5A CN202111390832A CN114037480A CN 114037480 A CN114037480 A CN 114037480A CN 202111390832 A CN202111390832 A CN 202111390832A CN 114037480 A CN114037480 A CN 114037480A
Authority
CN
China
Prior art keywords
city
new
charging
module
demand
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.)
Granted
Application number
CN202111390832.5A
Other languages
Chinese (zh)
Other versions
CN114037480B (en
Inventor
赵东
马华东
尤晓勇
王义总
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.)
Beijing University of Posts and Telecommunications
Original Assignee
Beijing University of Posts and Telecommunications
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 Beijing University of Posts and Telecommunications filed Critical Beijing University of Posts and Telecommunications
Priority to CN202111390832.5A priority Critical patent/CN114037480B/en
Priority claimed from CN202111390832.5A external-priority patent/CN114037480B/en
Publication of CN114037480A publication Critical patent/CN114037480A/en
Application granted granted Critical
Publication of CN114037480B publication Critical patent/CN114037480B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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
    • G06Q30/0202—Market predictions or forecasting for commercial activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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—Location or geographical consideration
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Development Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Software Systems (AREA)
  • Biomedical Technology (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Game Theory and Decision Science (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • Human Resources & Organizations (AREA)
  • Primary Health Care (AREA)
  • Tourism & Hospitality (AREA)
  • Charge And Discharge Circuits For Batteries Or The Like (AREA)

Abstract

The invention discloses a new-energy vehicle charging pile demand prediction and deployment optimization method for a new city, which uses a city migration learning paradigm, utilizes city migration knowledge with rich historical data to help a target city to predict user demands of a candidate deployment scheme, and simultaneously utilizes a heuristic thought to provide a new model of prediction-planning iterative interaction and a corresponding efficient solution method, quickly searches an optimal solution in an explosive scheme combination, and ensures the preference and convergence of the method. The method of the invention simultaneously carries out prediction and planning, ensures the optimization tendency and convergence of the deployment scheme adjusting process, improves the solving efficiency by a heuristic method, greatly reduces the time required by optimization, simultaneously uses the dynamic planning thought to search the optimal solution in the iterative process, and really realizes the synchronous demand prediction and the charging pile deployment scheme optimization.

Description

New-energy vehicle charging pile demand prediction and deployment optimization method for new city
Technical Field
The invention relates to the technical field of urban calculation, urban planning and new energy vehicles, in particular to a new city-oriented new energy vehicle charging pile demand prediction and deployment optimization method.
Background
The method for predicting the charging demand of the charging pile deployed in a new city is mainly divided into two types: prediction from implicit data and prediction from explicit data.
1. And predicting according to implicit data. The method mainly comprises the following steps: and modeling the association between the parking demand, population distribution, traffic flow, refueling record and the charging demand by using implicit data which is associated with the charging demand, such as the parking demand, population distribution, traffic flow, refueling record and the like, and further calculating the charging demand of the new city. The implicit data and the charging demand data have different natural attributes, which bring substantial space-time mobility differences, and the predicted charging demand is very easy to make mistakes.
2. And predicting according to the explicit data. The method mainly comprises the following steps: and calculating the charging requirement of the whole city through the acquired partial charging requirement data. For example, a charging trajectory is extracted from the traveling trajectories of a plurality of charging cars to obtain the distribution of charging demands; or the transaction data is crawled from the charging transaction platform, and point of interest information (POI) corresponding to the charging station is combined to serve as the charging requirement. However, in a completely new city, it is almost impossible to directly acquire such explicit data.
However, when a new city is faced, a charging vehicle needs to be popularized in the new city, and then a charging station needs to be deployed in the city, and we naturally expect that the deployment of the charging station can bring greater benefit, so we need to deploy according to the charging requirement of the city, but because the new city lacks historical data of the charging station, the charging requirement of the city cannot be predicted, and further an optimal deployment scheme cannot be selected from a candidate charging station set.
The definition of the charging pile deployment optimization is as follows: for a group of candidate schemes, the maximum benefit which can be obtained in each scheme is automatically calculated, and then the optimal solution is found from the whole candidate scheme set. From the optimization goal, the existing work mainly aims at maximizing the social benefits of the charging station, for example, meeting the charging requirements of the vehicle owners as much as possible and reducing the queuing waiting time.
The invention aims to solve the problem of predicting the charging demand and optimizing the deployment scheme under the condition of no charging data when a charging station is deployed in a new city.
Disclosure of Invention
Aiming at the problems, the invention provides a new energy vehicle charging pile demand prediction and deployment optimization method for a new city.
In order to achieve the above purpose, the invention provides the following technical scheme:
a new-energy-vehicle charging pile demand prediction and deployment optimization method for a new city comprises the following steps:
s1, firstly, initializing a charging pile distribution in the target city
Figure BDA0003368762310000021
S2, charging pile distribution according to source city
Figure BDA0003368762310000022
Target city charging pile distribution
Figure BDA0003368762310000023
Source city data DSCTarget city data DTCTraining a charging demand prediction model and predicting charging demand distribution of a target city
Figure BDA0003368762310000024
S3, calculating daily average revenue R of current charging pile distribution by using prediction resultTC；
S4, giving a revenue increase threshold value theta, if RTCTheta is less than or equal to theta, the process is finished, and the distribution of the charging piles in the current target city is returned
Figure BDA0003368762310000025
Otherwise, executing step S5;
s5, constructing candidate sets according to current charging pile distribution
Figure BDA0003368762310000026
S6, pair
Figure BDA0003368762310000027
Traversing, and predicting by using a charging demand prediction model to obtain a new charging demand;
s7, searching by dynamic planning thought
Figure BDA0003368762310000028
Optimal solution of (2), to
Figure BDA0003368762310000029
Carrying out primary updating; it jumps to step S2.
Further, the charging demand prediction model comprises a context module, a portrait feature module, a demand prediction module and a field self-adaption module: wherein the context characteristics of each charging station in the source city and the target city are input into the context module and output as context characteristics fc(ii) a The portrait characteristics of each charging station of the source city and the target city are input into the portrait characteristic module and outputp(ii) a Context feature fcAnd portrait features fpCombining the obtained site joint characteristics f, inputting the site joint characteristics f and the time intervals into a demand prediction module, and outputting the charging demand containing the charging demand in each time interval
Figure BDA00033687623100000210
Site association feature f input domain adaptive module, output domain tag
Figure BDA00033687623100000211
Indicating the domain to which the site feature belongs.
Further, the context module is composed of a convolution block, a spatiotemporal attention module, a convolution block and a global pooling layer in sequence, wherein each convolution block comprises a convolution layer, a normalization layer and a ReLU activation function.
Further, the image feature module is composed of two fully connected layers with a ReLU activation function.
Furthermore, the demand forecasting module consists of an embedding layer, two full-connection layers and a full-connection layer, the time interval is converted into an array q through the embedding layer, the site characteristics connect the output result with the q through the two full-connection layers, and then the output result is input into the full-connection layer to obtain the forecasting result.
Further, the demand prediction module loss function:
Ldemand＝(1-α)Lreg+αLrank
wherein L isregRepresents the mean square error, LrankIndicating a ranking error.
Further, LregThe calculation formula of (2) is as follows:
Figure BDA0003368762310000031
SSCindicating the charging requirements of the source city and,
Figure BDA0003368762310000032
and y represents the predicted and true values of the target city charging demand, respectively.
Further, LrankThe calculation formula of (2) is as follows:
Figure BDA0003368762310000033
wherein, PijRepresenting y in the truth of the target cityiRatio yjThe probability of the higher ranking is that,
Figure BDA0003368762310000034
indicating y in the predictioniRatio yjHigher probability of ranking, PijThe calculation formula of (2) is as follows:
Figure BDA0003368762310000035
oij＝yi-yj(yi＞yj)。
further, the domain adaptation module is composed of two fully connected layers.
Further, the loss function of the domain adaptation module is:
Figure BDA0003368762310000036
Figure BDA0003368762310000037
Figure BDA0003368762310000038
wherein: d is a domain label and the domain name is,
Figure BDA0003368762310000041
is a predicted domain label, S ═ Ssc∪STC，
Figure BDA0003368762310000042
And
Figure BDA0003368762310000043
is a weight parameter that is a function of,
Figure BDA0003368762310000044
and
Figure BDA0003368762310000045
is a deviation parameter.
Compared with the prior art, the invention has the beneficial effects that:
the new-energy-vehicle charging pile demand prediction and deployment optimization method for the new city, provided by the invention, uses a city migration learning paradigm, utilizes city migration knowledge with rich historical data to help a target city to predict user demands of a candidate deployment scheme, and simultaneously utilizes a heuristic thought to provide a new model of prediction-planning iterative interaction and a corresponding efficient solution method, so that an optimal solution is rapidly searched in an explosive scheme combination, and the trend and the convergence of the method are ensured. The method of the invention simultaneously carries out prediction and planning, ensures the optimization tendency and convergence of the deployment scheme adjusting process, improves the solving efficiency by a heuristic method, greatly reduces the time required by optimization, simultaneously uses the dynamic planning thought to search the optimal solution in the iterative process, and really realizes the synchronous demand prediction and the charging pile deployment scheme optimization.
Drawings
In order to more clearly illustrate the embodiments of the present application or technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present invention, and other drawings can be obtained by those skilled in the art according to the drawings.
Fig. 1 is a schematic flow diagram of a new energy vehicle charging pile demand prediction and deployment optimization method for a new city.
Detailed Description
For a better understanding of the present solution, the method of the present invention is described in detail below with reference to the accompanying drawings.
1. Problem scenario
Now, if a new city needs to be popularized, charging stations need to be deployed in the city. We naturally expect that the deployment of the charging stations can bring greater benefits, so we need to deploy according to the charging needs of the city, but because a new city lacks the historical data of the charging stations, the charging needs of the city cannot be predicted, and further, an optimal deployment scheme cannot be selected from a candidate charging station set.
2. Problem definition
We define: charging station
Figure BDA0003368762310000051
Wherein liIs ciThe geographical location of the mobile station (c),
Figure BDA0003368762310000052
and
Figure BDA0003368762310000053
are respectively ciThe number of the medium slow filling piles and the fast filling piles,
Figure BDA0003368762310000054
and
Figure BDA0003368762310000055
are respectively at ciThe cost of deploying one slow-fill pile and one fast-fill pile,
Figure BDA0003368762310000056
and
Figure BDA0003368762310000057
respectively represent CiThe service prices of the slow-fill pile and the fast-fill pile in the time interval T.
Figure BDA0003368762310000058
And
Figure BDA0003368762310000059
respectively represent ciThe utilization of the slow and fast charging piles during the time interval T is used in the present problem to represent the charging demand of the charging station.
We consider two cities: source city, charging stations have been deployed, and the set of charging stations is marked as CsC. Each deployed charging station c in the source cityi∈CSCIs known. Target city, no charging stations yet deployed, a candidate deployment partyRecord as CTC. Candidate charging stations c in the target cityi∈CTCOnly is provided with
Figure BDA00033687623100000510
These 5 attributes are known.
We use collections
Figure BDA00033687623100000511
To represent a target city deployment scenario CTCThe number of slow-charging and fast-charging piles of each station.
Figure BDA00033687623100000512
Is represented by CSCThe number of slow-charging and fast-charging piles of each station. DSCAnd DTCThe method is used for representing multi-source situation data in a source city and a target city respectively, namely POI data, road network data and traffic data.
We propose a charge demand prediction model f, given
Figure BDA00033687623100000513
DSCAnd DTCAnd charging demand of source city
Figure BDA00033687623100000514
Under the condition of (1), minimizing the predicted value of the charge demand
Figure BDA00033687623100000515
With the true value YTCThe error between, namely:
Figure BDA00033687623100000516
deployment scheme for solving target city on the basis
Figure BDA00033687623100000517
Given CTC,
Figure BDA00033687623100000518
YSC、DSC、DTCUnder the conditions of the charging demand prediction method f and the budget upper limit B, the revenue R is maximized and the deployment cost does not exceed the budget upper limit B, that is:
Figure BDA00033687623100000519
Figure BDA00033687623100000520
Figure BDA00033687623100000521
3. charging demand prediction model
The architecture of the charging demand prediction model is shown in fig. 1, and the charging demand prediction model is composed of four components, which are respectively: context module, portrait feature module, demand forecast module, domain self-adapting module.
1) Context module
Module input: all the context characteristics of each charging station in the source city and the target city include POI, road network characteristics, traffic characteristics of the charging station and its neighbor charging stations.
And (3) module output: context feature fc。
The module structure is as follows: volume block + spatio-temporal attention module + volume block + global pooling layer. Each convolution block contains one convolution layer, one normalization layer, and one ReLU activation function. The input for each volume block is
Figure BDA0003368762310000061
Then outputting:
Figure BDA0003368762310000062
wherein WcAnd bcAre all learnable parameters, representing convolution operations,BN stands for batch normalization. To avoid overfitting, the discard operation will be enabled after the first volume block. The spatial attention module is used to encode the spatial correlation. The second convolution module is used to improve performance. The global pooling layer is used to reduce feature dimensionality.
2) Portrait feature module
Module input: and (3) portrait characteristics of each charging station of the source city and the target city, namely charging pile information of the charging stations per se and the neighbor charging stations.
And (3) module output: image feature fp。
The module structure is as follows: two fully-connected layers with a ReLU activation function.
3) Demand forecasting module
Module input: site association feature f (by f)cAnd fpCombined) and time intervals.
And (3) module output: including the need for charging in each time interval
Figure BDA0003368762310000064
The module structure is as follows: the time interval is converted into an array q by an embedding layer. And the site characteristics connect the output result with q through two full-connection layers, and then input the output result into one full-connection layer to obtain a prediction result.
Loss function: l isdemand＝(1-α)Lreg+αLrank
Wherein L isregRepresenting mean square error, SSCIndicating the charging requirements of the source city and,
Figure BDA0003368762310000063
and y represents the predicted and true values of the target city charging demand, respectively.
Figure BDA0003368762310000071
LrankIndicating a ranking error. Wherein, PijRepresenting y in the truth of the target cityiRatio yjThe probability of the higher ranking is that,
Figure BDA0003368762310000072
indicating y in the predictioniRatio yjHigher probability of ranking.
oij＝yi-yj(yi＞yj)
Figure BDA0003368762310000073
Figure BDA0003368762310000074
4) Domain adaptive module
Module input: site association feature f (by f)cAnd fpCombine to get)
And (3) module output: domain label
Figure BDA0003368762310000075
For indicating the domain to which the site feature belongs
The module structure is as follows: two fully connected layers.
Loss function:
Figure BDA0003368762310000076
Figure BDA0003368762310000077
Figure BDA0003368762310000078
wherein: d is a domain label and the domain name is,
Figure BDA0003368762310000079
is a predicted domain label, S ═ Ssc∪STC，
Figure BDA00033687623100000710
And
Figure BDA00033687623100000711
is a weight parameter that is a function of,
Figure BDA00033687623100000712
and
Figure BDA00033687623100000713
is a deviation parameter.
For the entire charge demand prediction module, our goal is to minimize LregAnd LrankTo maximize Ldomain. The final joint loss function is:
L＝(1-α)Lreg+αLrank-βLdomain
4. charging pile deployment scheme optimization module
Module object: and in a given candidate charging station position set, the number of the fast-slow pile piles of each charging station is adjusted, and the daily income is maximized under the condition that the final total cost is not more than the upper limit B.
Module input: revenue growth threshold θ, total budget B, source city data DSCSource city charging pile distribution
Figure BDA00033687623100000714
Target city data DTC
And (3) module output: charging pile distribution of target city
Figure BDA00033687623100000715
Constructing a candidate set: if the number of the fast and slow charging piles of a certain charging station is (n)SF，nSs), defining a candidate set NiComprises the following steps:
Figure BDA0003368762310000081
charging pile distribution for target city
Figure BDA0003368762310000082
In the case of a composite material, for example,
Figure BDA0003368762310000083
the new-energy-vehicle charging pile demand prediction and deployment optimization method for the new city comprises the following steps:
(1) firstly, a charging pile distribution is initialized in a target city
Figure BDA0003368762310000084
(2) According to
Figure BDA0003368762310000085
DSC、DTCTraining a charging demand prediction model f and predicting the charging demand distribution of a target city
Figure BDA0003368762310000086
(3) Calculating daily average revenue R of current charging pile distribution by using prediction resultTC；
(4) If R isTCTheta is less than or equal to theta, the process is finished, and the distribution of the charging piles in the current target city is returned
Figure BDA0003368762310000087
Otherwise, executing the step (5);
(5) constructing candidate sets according to current charging pile distribution
Figure BDA0003368762310000088
(6) To pair
Figure BDA0003368762310000089
Traversing, and obtaining a new charging demand by using f prediction;
(7) looking up by dynamically planned thought
Figure BDA00033687623100000810
Optimal solution of (2), to
Figure BDA00033687623100000811
An update is performed. And (4) jumping to the step (2).
The method realizes the prediction of the charging demand of the target city by using the historical data of the public charging station of the source city. The implicit data set is used for calculating the charging requirement of the target city, the intrinsic space-time mobility difference is brought by the difference of natural attributes among data, and the predicted charging requirement is very prone to errors. For new cities, explicit data is difficult to collect. According to the method and the device, charging data of the source city are used, multi-source data such as POI of the source city and POI of the target city are combined, image characteristics and situation characteristics which can be distinguished are learned, and then the charging requirement of the target city is predicted. Meanwhile, a domain adaptive network is introduced to guide the network to learn high-level semantic features irrelevant to the city. In the aspect of model design, the space-time characteristics of the charging behaviors are kept, meanwhile, the influence of the road network information, POI information, traffic information and other context characteristics on the charging behaviors is considered, and the mutual influence among the charging stations is introduced. Meanwhile, a domain self-adaptive network is introduced aiming at the domain deviation problem in the urban transfer learning, so that the distribution deviation of an input space caused by the change on an observation system is reduced. Therefore, the charging demand prediction method and the charging demand prediction device achieve the charging demand prediction under the condition that no target city charging data exists.
The method of the invention simultaneously carries out prediction and planning, ensures the optimization tendency and convergence of the deployment scheme adjusting process, and improves the solving efficiency by a heuristic method. In one iteration of the optimized deployment scenario, there are 5 candidate scenarios per site, which results in a total
Figure BDA00033687623100000812
New deployment scenarios, where training and prediction are individually performed for each deployment scenario, are unacceptable. The application only carries out 1 training and 5| C in one iteration processTCPrediction of degree I, greatly reducingThe time required by optimization is reduced, meanwhile, the optimal solution in the iterative process is searched by using a dynamic planning idea, and the synchronous demand prediction and the optimization of the charging pile deployment scheme are really realized.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: it is to be understood that modifications may be made to the technical solutions described in the foregoing embodiments, or equivalents may be substituted for some of the technical features thereof, but such modifications or substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (10)

1. A new-energy-vehicle charging pile demand prediction and deployment optimization method for a new city is characterized by comprising the following steps:
s1, firstly, initializing a charging pile distribution in the target city
Figure FDA0003368762300000011
S2, charging pile distribution according to source city
Figure FDA0003368762300000012
Target city charging pile distribution
Figure FDA0003368762300000013
Source city data DSCTarget city data DTCTraining a charging demand prediction model and predicting charging demand distribution of a target city
Figure FDA0003368762300000014
S3, calculating daily average revenue R of current charging pile distribution by using prediction resultTC；
S4, giving a revenue increase threshold value theta, if RTCTheta is less than or equal to theta, the process is finished, and the distribution of the charging piles in the current target city is returned
Figure FDA0003368762300000015
Otherwise, executing step S5;
s5, constructing candidate sets according to current charging pile distribution
Figure FDA0003368762300000016
S6, pair
Figure FDA0003368762300000017
Traversing, and predicting by using a charging demand prediction model to obtain a new charging demand;
s7, searching by dynamic planning thought
Figure FDA0003368762300000018
Optimal solution of (2), to
Figure FDA0003368762300000019
Carrying out primary updating; it jumps to step S2.
2. The new-city-oriented new-energy-vehicle charging pile demand prediction and deployment optimization method according to claim 1, wherein the charging demand prediction model comprises a context module, a portrait feature module, a demand prediction module and a field self-adaption module: wherein the context characteristics of each charging station in the source city and the target city are input into the context module and output as context characteristics fc(ii) a The portrait characteristics of each charging station of the source city and the target city are input into the portrait characteristic module and outputp(ii) a Context feature fcAnd portrait features fpCombining the obtained site joint characteristics f, inputting the site joint characteristics f and the time intervals into a demand prediction module, and outputting the charging demand containing the charging demand in each time interval
Figure FDA00033687623000000110
The site association feature f is input to a domain adaptation module,output domain label
Figure FDA00033687623000000111
Indicating the domain to which the site feature belongs.
3. The new-energy-vehicle charging-pile demand prediction and deployment optimization method for the new city according to claim 2, wherein the context module sequentially comprises a convolution block, a spatio-temporal attention module, a convolution block and a global pooling layer, and each convolution block comprises a convolution layer, a normalization layer and a ReLU activation function.
4. The new-city-oriented new-energy-vehicle charging pile demand prediction and deployment optimization method according to claim 2, wherein the image feature module is composed of two full-connection layers with a ReLU activation function.
5. The new-energy-vehicle-charging-pile demand forecasting and deployment optimizing method for the new city according to claim 2, wherein the demand forecasting module is composed of an embedded layer, two fully-connected layers and a fully-connected layer, a time interval is converted into an array q through the embedded layer, site characteristics are connected with the q through the two fully-connected layers, and then the output result is input into the fully-connected layer to obtain a forecasting result.
6. The new-city-oriented new-energy vehicle charging pile demand forecasting and deployment optimizing method according to claim 2, characterized in that a demand forecasting module loss function:
Ldemand＝(1-α)Lreg+αLrank
wherein L isregRepresents the mean square error, LrankIndicating a ranking error.
7. The new city-oriented new energy vehicle charging pile demand prediction and deployment optimization method according to claim 6, wherein L is LregThe calculation formula of (2) is as follows:
Figure FDA0003368762300000021
SSCindicating the charging requirements of the source city and,
Figure FDA0003368762300000022
and y represents the predicted and true values of the target city charging demand, respectively.
8. The new city-oriented new energy vehicle charging pile demand prediction and deployment optimization method according to claim 6, wherein L is LrankThe calculation formula of (2) is as follows:
Figure FDA0003368762300000023
wherein, PijRepresenting y in the truth of the target cityiRatio yjThe probability of the higher ranking is that,
Figure FDA0003368762300000024
indicating y in the predictioniRatio yjHigher probability of ranking, PijThe calculation formula of (2) is as follows:
Figure FDA0003368762300000025
oij＝yi-yj(yi＞yj)。
9. the new-city-oriented new-energy-vehicle charging pile demand prediction and deployment optimization method according to claim 2, wherein the domain self-adaptation module is composed of two full-connection layers.
10. The new-city-oriented new-energy-vehicle charging pile demand prediction and deployment optimization method according to claim 2, wherein the loss function of the field adaptive module is as follows:
Figure FDA0003368762300000026
Figure FDA0003368762300000031
Figure FDA0003368762300000032
wherein: d is a domain label and the domain name is,
Figure FDA0003368762300000033
is a predicted domain label, S ═ Ssc∪STC，
Figure FDA0003368762300000034
And
Figure FDA0003368762300000035
is a weight parameter that is a function of,
Figure FDA0003368762300000036
and
Figure FDA0003368762300000037
is a deviation parameter.
CN202111390832.5A 2021-11-23 New energy vehicle charging pile demand prediction and deployment optimization method for new city Active CN114037480B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111390832.5A CN114037480B (en) 2021-11-23 New energy vehicle charging pile demand prediction and deployment optimization method for new city

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111390832.5A CN114037480B (en) 2021-11-23 New energy vehicle charging pile demand prediction and deployment optimization method for new city

Publications (2)

Publication Number Publication Date
CN114037480A true CN114037480A (en) 2022-02-11
CN114037480B CN114037480B (en) 2024-06-25

Family

ID=

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116629413A (en) * 2023-05-10 2023-08-22 金陵科技学院 Global optimal automatic deployment method for mobile charging device based on random forest and Dijkstra

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103793758A (en) * 2014-01-23 2014-05-14 华北电力大学 Multi-objective optimization scheduling method for electric vehicle charging station including photovoltaic power generation system
CN106777835A (en) * 2017-02-15 2017-05-31 国网江苏省电力公司苏州供电公司 The city electric car charging network planing method of multiple optimization aims is considered simultaneously
US20190340543A1 (en) * 2018-05-01 2019-11-07 Fleetonomy, Ltd. System and method for optimizing vehicle fleet deployment
CN111311122A (en) * 2020-03-23 2020-06-19 广东工业大学 Charging station construction method for smart city

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103793758A (en) * 2014-01-23 2014-05-14 华北电力大学 Multi-objective optimization scheduling method for electric vehicle charging station including photovoltaic power generation system
CN106777835A (en) * 2017-02-15 2017-05-31 国网江苏省电力公司苏州供电公司 The city electric car charging network planing method of multiple optimization aims is considered simultaneously
US20190340543A1 (en) * 2018-05-01 2019-11-07 Fleetonomy, Ltd. System and method for optimizing vehicle fleet deployment
CN111311122A (en) * 2020-03-23 2020-06-19 广东工业大学 Charging station construction method for smart city

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116629413A (en) * 2023-05-10 2023-08-22 金陵科技学院 Global optimal automatic deployment method for mobile charging device based on random forest and Dijkstra

Similar Documents

Publication Publication Date Title
CN112016812B (en) Multi-unmanned aerial vehicle task scheduling method, system and storage medium
AU2020102350A4 (en) A Spark-Based Deep Learning Method for Data-Driven Traffic Flow Forecasting
CN110222907B (en) Electric vehicle charging station planning method and terminal equipment
CN109960544B (en) Task parallel scheduling method based on data driving type agile satellite
CN113485429B (en) Route optimization method and device for air-ground cooperative traffic inspection
CN109087509B (en) Road network traffic running state prediction method
CN107180274B (en) Typical scene selection and optimization method for electric vehicle charging facility planning
An et al. Optimal scheduling of electric vehicle charging operations considering real-time traffic condition and travel distance
Kim et al. Idle vehicle relocation strategy through deep learning for shared autonomous electric vehicle system optimization
CN115547052B (en) Dynamic demand response electric bus scheduling method for improving self-adaptive large neighborhood algorithm
Gao et al. Optimal multi-taxi dispatch for mobile taxi-hailing systems
CN111898793B (en) Path selection method considering user perception difference in combined travel mode
CN114037175B (en) Large-scale public transportation network hierarchical optimization method based on multi-scale clustering
CN109063759A (en) A kind of neural network structure searching method applied to the more attribute forecasts of picture
CN115545758A (en) Method and system for self-adaptive incremental site selection of urban service facilities
Shi et al. Deep q-network based route scheduling for transportation network company vehicles
CN114092176A (en) Urban commuting regular bus planning method based on bus
CN112381398A (en) Electric vehicle charging station site selection method and system considering limited trip of user
CN115966083A (en) Distributed traffic flow prediction method and system based on time-space decoupling
CN112036598A (en) Charging pile use information prediction method based on multi-information coupling
CN114969367A (en) Cross-language entity alignment method based on multi-aspect subtask interaction
CN114117910A (en) Electric vehicle charging guide strategy method based on layered deep reinforcement learning
Chen et al. Finding next high-quality passenger based on spatio-temporal big data
Wang et al. Double graph attention actor-critic framework for urban bus-pooling system
CN111091286B (en) Public bicycle scheduling method

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