CN114742340A

CN114742340A - Optimal layout solving method for intelligent network connection sharing electric vehicle charging station in large-scale road network

Info

Publication number: CN114742340A
Application number: CN202210054494.6A
Authority: CN
Inventors: 尚夏; 戴钰超; 章立辉; 王亦兵
Original assignee: Zhejiang University ZJU
Current assignee: Zhejiang University ZJU
Priority date: 2022-01-18
Filing date: 2022-01-18
Publication date: 2022-07-12
Anticipated expiration: 2042-01-18
Also published as: CN114742340B

Abstract

The invention belongs to the technical field of intelligent traffic, and discloses an optimal layout solving method for intelligent networking shared electric vehicle charging stations in a large-scale road network, which comprises the following steps of 1: establishing a set of charging station layout mixed integer linear programming model aiming at minimizing the total system operation cost; 2: solving the established model by combining a time window rolling staged dynamic programming method and a genetic algorithm; the upper layer of the invention uses a genetic algorithm to generate an initial population comprising a number of individuals, wherein each individual represents a charging station layout scheme. And (4) according to the individual fitness evaluation value provided by the lower layer problem, carrying out genetic evolution operations such as selection, hybridization and mutation on each generation of population, and forming a new generation of population after reinsertion. The iterative evolution is carried out until the minimum value of the individual fitness in the population reaches convergence. The original large-scale problem is split in the time dimension by a method of continuously rolling in the time stage, and the high dimension is reduced to the low dimension, so that the original problem is effectively solved.

Description

Optimal layout solving method for intelligent networking sharing electric vehicle charging station in large-scale road network

Technical Field

The invention belongs to the technical field of intelligent traffic, and particularly relates to an optimal layout solving method for intelligent network connection sharing electric vehicle charging stations in a large-scale road network.

Background

From the perspective of the system, intelligent network sharing of vehicles is expected to greatly improve traffic travel, for example, the average vehicle retention rate can be reduced, traffic congestion is relieved, the population coverage of traffic service is enlarged, the vehicle utilization rate is improved, the land occupancy rate is reduced, and the like; from the perspective of travelers, the internet sharing of vehicles can reduce the capital cost of purchasing vehicles and the cost of maintaining vehicles, but can also share the travel comfort level similar to private cars, and for travelers with short mileage, the automobile sharing is a lower-cost option than the possession of private cars and is more flexible than public transport services. Therefore, the sharing of the intelligent networking vehicles has remarkable social benefits.

On the other hand, motorization is an important direction of automobile evolution. However, the electric vehicle has many limitations in terms of limited driving range, insufficient charging facilities, long charging time, and high upfront purchase cost. The problems of shortage of charging facilities and unreasonable layout are one of the important factors for restricting the rapid development of electric automobiles. In the intelligent internet sharing system, each electric vehicle needs to continuously complete continuous travel service, and the electric quantity is continuously consumed in the process, so the arrangement problem of the charging facility is particularly important. In an intelligent internet shared traffic system, how to optimize and arrange charging facilities needs to be researched in advance. The scales of different intelligent networking sharing fleets, different sharing electric vehicle travel plans and the like correspond to different electric quantity consumption processes, so the factors are comprehensively considered when the network layout and the charging pile capacity of future charging stations are optimized.

In China, the scale of a general urban road network is large, the number of traffic network nodes is large, the trip demand of passengers is large, the dimension of a charging station optimization problem is too high, the scale is too large, a large amount of memory and time are consumed when a computer solves a layout scheme, and the problem is difficult to directly solve by using an existing commercial solver, so that an effective solving method suitable for a large-scale road network needs to be developed.

Disclosure of Invention

The invention aims to provide an optimized layout solving method for intelligent network connection sharing electric vehicle charging stations in a large-scale road network, which aims to solve the technical problems that the dimension of the charging station optimization problem is too high, the scale is too large, a large amount of memory and time are consumed when a computer solves the layout scheme, and the direct solution is difficult to realize by using the conventional commercial solver.

In order to solve the technical problems, the specific technical scheme of the optimal layout solving method for the intelligent network connection sharing electric vehicle charging station in the large-scale road network is as follows:

an optimal layout solving method for intelligent network connection sharing electric vehicle charging stations in a large-scale network comprises the following steps:

c1: establishing a set of charging station distribution mixed integer linear programming Model (MILP) aiming at minimizing the total system operation cost;

c2: solving the established model by combining a time window rolling staged dynamic programming method and a genetic algorithm; the upper layer generates an initial population containing a plurality of individuals by using a genetic algorithm, wherein each individual represents a charging station layout scheme; according to the individual fitness evaluation value provided by the lower-layer problem, each generation of population is subjected to genetic evolution operations of selection, hybridization and mutation, and is reinserted to form a new generation of population; performing iterative evolution until the minimum value of the individual fitness in the population converges; the charging station layout scheme represented by the individual with the minimum fitness value is the optimal layout;

the lower layer applies a staged dynamic programming method of time window rolling to solve a mixed integer linear programming Model (MILP) after an individual layout scheme is given, so as to obtain the fitness of each individual; under each individual scheme, dividing a time period covered by a Model (MILP) into a plurality of small stages, and dividing large-scale passenger trip data into a plurality of small data according to each time stage; in each stage, solving a mixed integer linear programming Model (MILP) with reduced scale according to the current vehicle state of a road network and the travel demand information of passengers in the stage, and after solving in each stage, correspondingly updating vehicle path scheduling and charging decisions and state information of all vehicles to serve as an initial value of solving in the next stage; in the same way, the original large-scale problem is split in the time dimension by a method of continuously rolling in the time stage, and the high dimension is reduced to the low dimension; calculating and adding the obtained objective function values of all stages to obtain a fitness value of each population individual, wherein the fitness value is the total cost of the charging station layout strategy represented by the individual;

c3: and summarizing a large-scale road network solving method.

Further, the mixed integer linear programming Model (MILP) comprises a traffic network module, an objective function module, an electric vehicle passenger receiving and sending path module, a travel time module, an electric vehicle queuing waiting time module, an electric vehicle electric quantity module and a charging station module;

c1.1: building a traffic network module: the optional starting point of the electric vehicle is represented by a parameter O, the set of the boarding points of all passengers is represented by P, the set of the disembarking points of all passengers is represented by D, the set of the alternative point numbers of all charging stations is represented by U,

the maximum number of charging piles which can be installed at the alternative points of the charging station n is represented; v_maxRepresenting the number of electric vehicles which are allowed to be released in the network at most; c_maxIndicating that a limit on the number of charging stations allowed to be established in the road network is set; c_nRepresenting a set of all alternative charging piles of the n charging station alternative points; q_maxShowing the quantity limit of the charging piles put in a road network; c represents all charging station charging pile alternative point sets; t represents a return point of the electric vehicle, which is a virtual terminal point, and the distance from the electric vehicle to each other node is 0; n represents a set of all nodes; v represents the set of all the alternative electric vehicles, and a digital model of a physical traffic network is established;

c1.2: building an objective function module: by a parameter c_t、c_w、c_e、c_q、c_f、c_r、t_ij、

a_i、E、

r_jCalculating a minimum value of the sum of total costs of the respective stages; i.e., a minimum value of a sum of a travel time cost of the electric vehicle, a use cost of the autonomous electric vehicle and the charging post, a charging fee of the electric vehicle at the charging station, a waiting time cost of the electric vehicle at the charging station, and a penalty cost due to the node not being reached on time, wherein c_tRepresenting the time cost of the electric vehicle running for unit time; c. C_wThe penalty cost of unit waiting time for waiting at the getting-on point of the passenger due to the delay of the electric vehicle is shown; c. C_eRepresents the cost of energy consumption per unit of electricity; c. C_qThe unit queuing time cost of the electric vehicle in the charging pile queuing is shown; c. C_fRepresents the purchase and use cost of each electric vehicle; c. C_rThe construction and use cost of each charging pile is represented; t is t_ijRepresenting the travel time from node i e N to node j e N;

representing the time when the electric vehicle V belongs to V and actually reaches the node i belongs to N;

the time that the electric vehicle V belongs to V and starts to charge in the charging pile i belongs to C is represented;

the time of the electric vehicle V belonging to V from the passenger getting-on point i belonging to P is represented; a is_iRepresenting the time for the passenger to reach the boarding point i epsilon P; e represents the maximum electric quantity of the electric vehicle;

representing the electric quantity when the electric vehicle V belongs to V and reaches the node i belongs to N;

representing a variable of 0-1, if the electric vehicle V belongs to V and directly reaches a node j belongs to N from a node i belongs to N, and the value is equal to 1; otherwise, equal to 0; r is_jRepresenting a variable of 0-1, and if a charging pile candidate point j belongs to C, installing a charging pile, wherein the variable is equal to 1; otherwise, it equals 0.

C1.3: the electric vehicle receives and sends the building of the passenger path module: and (3) generating a passenger receiving and sending path of the shared electric vehicle through formula calculation constraint, and enabling the vehicle to complete the passenger receiving and sending tasks of each stage according to a specified path rule: at each stage, the electric vehicle v starts from a departure point O, goes to a passenger boarding point P to pick up passengers, and conveys the passengers to a corresponding passenger alighting point D; if the electric quantity v of the electric vehicle is not enough to complete the next task of receiving and sending the passengers, the electric vehicle goes to a charging pile C for charging, then goes to other passengers to a boarding point P for receiving the passengers, and so on; when the electric vehicle v finishes all the passenger receiving and sending demands in the stage, the electric vehicle v stays at a passenger getting-off point D, and the electric vehicle v is guided to return to a virtual terminal T to represent that the electric vehicle v finishes the passenger receiving and sending task in the stage; the passenger getting-off point D where the electric vehicle v stays at the last stage is the starting point O of the next stage;

c1.4, building a travel time module: by the parameters g, v _ start, v₀_charged、s₀、c₀、 y₀Charging, calculating the travel time of the automatic driving electric automobile at each node; wherein g represents the charging efficiency of the charging pile, i.e., the amount of charge per unit time; v _ start represents the departure time of all electric vehicles leaving the departure point before tracking and recording each stage, and v _ start of all electric vehicles is initialized to be 0; v. of₀The charge represents the number of the electric vehicle which is charged each time in each stage and is initialized to be an empty set; s is₀Showing and recording electric vehicle v in each stage₀Initializing the charging to be an empty set at the time when the charging pile starts to charge; c. C₀Showing and recording electric vehicle v in each stage₀A charging pile number corresponding to charging is initialized to be an empty set;y₀"charged" means to record the electric vehicle v in each stage₀Initializing the electric quantity when the charge reaches the charging pile to be an empty set;

c1.5: the electric vehicle queuing waiting time module is built: by the parameters described under C1.1-C1.4 and the parameters

And

calculating the queuing waiting time of the electric vehicle at each charging station; wherein

Is a variable from 0 to 1;

c1.6: electric vehicle electric quantity module is built: by the parameters described under C1.1-C1.4 and the parameters v _ energy, d_ijAnd h, describing the electricity consumption and the charging condition of the shared electric vehicle in the operation process, wherein v _ energy represents the electricity quantity of all the electric vehicles at the starting point before each stage of tracking and recording, and v _ energy of all the electric vehicles is initialized to E and d_ijThe shortest distance between the node i belonging to N and the node j belonging to N is represented, and h represents the electric quantity consumption rate of the electric vehicle, namely the electric quantity consumed by the electric vehicle in unit time; c1.7: building a charging station module: passing parameters

And determining the charging condition of the shared electric vehicle in the charging pile.

Further, the step C2 includes the following steps:

c2.1: individual coding;

c2.2: the iteration is performed using an evolutionary operation.

Further, step C2.1 is characterized by comprising the steps of: each individual is randomly coded by using a binary coding rule, 0 represents that no charging station or charging pile is built, 1 represents that a charging station or charging pile is built, and a variable Z representing whether a charging station is built or not is adopted_nAnd a variable r indicating whether the charging pile is installed or not_jWill result from generatingThe individual code value of (2).

Further, the step C2.2 performs genetic operations of selection, hybridization, and mutation on each generation of population, and forms a new generation of population after reinsertion, and the iterative evolution is performed until the minimum value of the fitness of the individuals in the population converges, and the charging station layout scheme represented by the individual with the minimum fitness value is the optimal layout.

Further, the step 3 comprises the following specific steps:

c3.1: initializing;

c3.2: individual coding;

c3.2: defining an evaluation function, and calculating the fitness value of each individual;

c3.4: performing population iteration by using a genetic algorithm;

c3.5: and obtaining an optimal charging station layout scheme.

Further, the C3.1 comprises the following specific steps:

c 3.1.1: introduction and study of road network:

obtaining the travel demand OD pairs of all the passengers in m hours in a road network and the departure time of the passenger travel demand OD pairs, the departure points or the return points of the electric vehicle, the nodes such as all the alternative points of the charging stations and the like, and the shortest paths between the nodes, and calculating to obtain the layout schemes of all the charging stations;

c3.1.2: a splitting stage:

dividing m hours into s stages, wherein the time of each stage is t seconds, and dividing all passenger travel demand OD pairs and departure time thereof into s parts according to the time stage of the departure time of the passenger;

c3.1.3: initializing a layout scheme of a charging station:

and randomly generating an initial charging station layout scheme.

Further, the C3.2 comprises the following specific steps:

c3.2.1: coding of charging station layout:

encoding the initial charging station layout scheme;

c3.2.2: the code that electric pile was laid:

and according to the layout condition of the initial charging station, coding the layout of the initial charging piles.

Further, the C3.3 adopts the established mixed integer linear programming model with the total cost minimized as the target as the evaluation function of each individual fitness, and includes the following specific steps:

and (3) circularly executing the following steps for each individual u in the initial population:

c3.3.1: initializing a relevant set; initializing the set v _ origin, v _ start, v _ energy, v _ terminal, v _ origin₀_charged，s₀，c₀，y₀A _ charged; establishing an aggregate cost to record the total cost of the electric vehicle for completing all passenger pick-up and delivery requirement tasks under the charging station layout scheme u, and initializing the cost to be 0;

c3.3.2: calculating a target function value of the electric vehicle for receiving and sending the passengers in stages;

each of the split s stages is k, and the following steps are executed in a circulating manner:

c3.3.2.1: building a traffic network; importing the OD pairs of the travel demands of the passengers in the k stage and the updated v _ origin set, constructing sets O, P, D, C, T and N according to 1.1-section traffic network construction rules, and updating the shortest paths of the nodes of the constructed traffic network;

c3.3.2.2: optimizing the implementation stage; solving a mixed integer linear programming Model (MILP) of the stage to obtain a vehicle passenger receiving and sending path strategy, an electric vehicle scale, a charging strategy and a target function value k _ cost which are optimal in the stage k, and updating the cost as cost + k _ cost;

c3.3.2.3: acquiring information of the vehicle running at the stage k; acquiring vehicles running at the stage k, starting points of running the vehicles, end points of running the vehicles, electric quantity of the running vehicles reaching the end points, time of the running vehicles reaching the end points and charging piles charged by the running vehicles, and respectively importing corresponding information into a set v₀, v₀_origin,v₀_destin,y₀_destin,tao₀_destin,c₀Used;

c3.3.2.4: updating the departure points of all the electric vehicles; updating the terminal point of the vehicle operated in the stage k into a set v _ origin by taking the terminal point of the vehicle operated in the stage k as the starting point of the vehicle operated in the next stage k + 1;

c3.3.2.5: updating the energy of all the electric vehicles at the starting point; taking the electric quantity of the vehicle running at the stage k reaching the terminal as the initial electric quantity of the vehicle running at the next stage k + 1;

c3.3.2.6: updating the departure time of the electric vehicle; the departure time v _ start of all vehicles in the next stage k +1 is updated to t (k +1), and the time tao when the running vehicle reaches the end point is further updated₀Compare _destinto t · (k + 1); if tao₀α destin > t · (k + 1): operating the vehicle v in phase k₀Time to endpoint tao₀Update of next phase k +1 operating vehicle v₀In the departure time v _ start; otherwise, v _ start remains unchanged;

c3.3.2.7: removing the records of the electric vehicle which finishes charging in the stage k; the electric vehicle is in a charging pile c₀Time s for used to end charging₀+(E-y₀) Comparison of/g with the start time t of stage k +1 (k + 1): if s is₀+(E-y₀) T (k + 1): delete electric vehicle at v₀_charged，s₀，c₀，y₀All records in _ charged; otherwise, these records are reserved;

c3.3.3: calculating a target function value of the electric vehicle for receiving and sending the passengers in stages;

c3.3.3.1: updating the traffic network; importing an updated v _ origin set, constructing sets of nodes in O, C, T and N according to a traffic network construction rule, and updating the shortest path of each node of the constructed traffic network;

c3.3.3.2: optimizing; calculating an optimal path and a charging strategy of the running electric vehicle to a return point and an optimal objective function value t _ cost, updating the cost as cost + t _ cost, wherein the cost value is a fitness value corresponding to the individual u;

wherein:

v _ origin represents the starting point numbers of all electric vehicles before each stage;

v _ terminal represents a return point for recording final return of all electric vehicles after all passenger travel requirement tasks are executed, and the return point is equal to the departure point of all electric vehicles provided in the import road network;

v₀representing all vehicles that are logged as operating in each phase;

v₀origin indicates the starting point of all vehicles recorded to be operated in each phase;

v₀"destin" indicates the end point of all vehicles that are operating in each phase is recorded;

y₀the step _destinrepresents the recording of the electric quantity of all running vehicles reaching the terminal point in each stage;

tao₀"destin" means the time to record the arrival of all vehicles at the end point in each phase;

c₀used represents the charging post that records the charging of all operating vehicles in each phase.

Further, the C3.4 sets the number of iterations gen of the population, and starts iteration from the second generation, and includes the following specific steps:

c3.4.1: carrying out hybridization selection; selecting w individuals with the best fitness from the parent population p, and copying one part of the individuals for cross use;

c3.4.2: performing cross evolution; crossing every two selected individuals in a uniform crossing mode, and deleting the fitness of the changed individuals;

c3.4.3: carrying out variant evolution; and (5) carrying out mutation on the crossed individuals in a disorder mutation mode.

Recalculating the fitness of the changed individual;

c3.4.4: carrying out reinsertion: selecting w individuals with the best fitness in the breeding offspring, inserting the w individuals into the parent, removing the individuals with the same codes, and taking the obtained result as the parent population for next iteration;

the C3.5 specific steps are as follows: if the minimum value of the individual fitness in the population converges after iteration, the individual code with the minimum fitness in the last generation of population is considered as the optimal layout scheme, the layout scheme is output, and if the minimum value of the individual fitness in the population does not reach the convergence, C3.5 is repeated until the convergence is reached.

The optimal layout solving method for the intelligent networking sharing electric vehicle charging station in the large-scale road network has the following advantages: the mixed integer linear programming model established by the invention has overlarge scale in a large-scale traffic network, and a computer needs to consume a large amount of memory and time during processing, so that the existing solver is difficult to directly solve. The invention combines a time window rolling staged dynamic programming method and a genetic algorithm to develop a set of solution method with a double-layer structure to solve the large-scale model. The upper layer uses a genetic algorithm to generate an initial population comprising a number of individuals, wherein each individual represents a charging station layout plan. According to the individual fitness evaluation value provided by the lower-layer problem, genetic evolution operations such as selection, hybridization, mutation and the like are carried out on each generation of population, and after reinsertion, a new generation of population is formed. The iterative evolution is carried out until the minimum value of the individual fitness in the population reaches convergence. The charging station layout scheme represented by the individual with the minimum fitness value is the optimal layout. And the lower layer solves the mixed integer linear programming model after the individual layout scheme is given by using a time window rolling staged dynamic programming method, so that the fitness of each individual is obtained. Under each individual scheme, a time period (usually one day or half day) covered by the model is divided into a plurality of small stages, and in each stage, the mixed integer linear programming model with reduced scale is solved according to the current vehicle states of the road network and the passenger travel requirement information in the stage. After the solution of each stage is finished, correspondingly updating the vehicle path scheduling and charging decision and the state information of all the vehicles, and taking the state information as the initial value of the solution of the next stage. By analogy, the original large-scale problem is split in the time dimension by a method of continuously rolling in the time stage, and the high dimension is reduced to the low dimension, so that the original problem is effectively solved.

Drawings

FIG. 1 is a schematic diagram of a passenger pick-up and delivery full-course path rule of an automatically driven shared electric vehicle;

FIG. 2 is a schematic diagram of a situation in which the actual queued charging situation is not met;

FIG. 3 is a schematic diagram illustrating a rule of a stepped passenger receiving and sending path of an autonomous driving shared electric vehicle;

FIG. 4 is a schematic diagram of a Chicago road network topology and an electric vehicle departure point and charging station alternate point layout;

FIG. 5 is a diagram illustrating an iterative convergence of population fitness minimum;

fig. 6 is a schematic diagram of optimal layout of a chicago network charging station when an electric vehicle completes 300 travel demand receiving and sending tasks.

Detailed Description

In order to better understand the purpose, structure and function of the present invention, the following describes in detail an optimized layout solution method for an intelligent networking sharing electric vehicle charging station in a large-scale road network with reference to the accompanying drawings.

The invention relates to an optimized layout solving method of intelligent networking sharing electric vehicle charging stations in a large-scale road network.

In the solution method, the upper layer generates an initial population including a plurality of individuals, each of which represents a charging station layout scheme, using a genetic algorithm. According to the individual fitness evaluation value provided by the lower-layer problem, genetic evolution operations such as selection, hybridization, mutation and the like are carried out on each generation of population, and after reinsertion, a new generation of population is formed. The iterative evolution is carried out until the minimum value of the individual fitness in the population reaches convergence. The charging station layout scheme represented by the individual with the minimum fitness value is the optimal layout.

And the lower layer solves a mixed integer linear programming Model (MILP) after the individual layout scheme is given by using a time window rolling staged dynamic programming method, so that the fitness of each individual is obtained. Under each individual scenario, the time period (typically one or half of a day) covered by the Model (MILP) is divided into a plurality of small phases, while large-scale passenger travel data is divided into several small pieces of data according to each time phase. And in each stage, solving a mixed integer linear programming Model (MILP) with reduced scale according to the current vehicle states of the road network and the passenger travel demand information in the stage. After the solution of each stage is finished, correspondingly updating the vehicle path scheduling and charging decision and the state information of all the vehicles, and taking the state information as the initial value of the solution of the next stage. By analogy, the original large-scale problem is split in the time dimension by a method of continuously rolling in the time stage, and the high dimension is reduced to the low dimension. And adding the objective function values obtained in each stage by calculation to obtain the fitness value of each population individual, wherein the fitness value is the total cost of the charging station layout strategy represented by the individual.

The method comprises the following implementation steps:

fitness evaluation function of individual charging station layout scheme-mixed integer linear programming Model (MILP)

1.1 sharing electric vehicle staged passenger receiving and sending model

1.1.1 traffic network set-up

The digitization of the physical traffic road network is realized through the following settings:

o: an alternative starting point for an electric vehicle, O ═ 1,2, 3., σ };

p: if δ boarding points are set for the set of boarding points of all passengers, P ═ σ +1, σ +2, σ + 3., σ + δ };

d: a set of all passenger drop-off points, D ═ σ + δ +1, σ + δ +2, σ + δ +3,.., σ +2 δ };

u: a set of all charging station candidate point numbers, wherein n charging station candidate points are set, and if U is {1, 2.., n };

the maximum number of charging piles which can be installed at the alternative points of the charging station n;

V_max: the number of electric vehicles which are allowed to be thrown into the network at most;

C_max: allowing a charging station number limit to be established in the road network;

C_n: set of all alternative charging piles at alternative points of charging station number n, e.g.

Q_max: limiting the number of charging piles thrown into a road network;

c: charging pile alternative point collection of all charging stations

C＝C₁∪C₂∪.....∪C_n＝{σ+2δ+1,σ+2δ+2,σ+2δ+3,...,σ+2δ+Q_max}；

T: return point of electric vehicle, T ═ σ +2 δ + Q_max+1}. It is a virtual end point, and the distance to each other node is 0;

n: a collection of all nodes, i.e. N ═ O ═ P ═ D ═ C ═ T;

v: set of all candidate electric vehicles, V ═ 1,2,3_max}。

1.1.2 objective function

The sum of total cost (including the running time cost of the electric vehicle, the use cost of the automatic driving electric vehicle and the charging pile, the charging expense of the electric vehicle at the charging station, the waiting time cost of the electric vehicle at the charging station, and the punishment cost caused by the fact that the electric vehicle cannot arrive at the node on time) of each stage is minimized through the following expression:

in the formula c_t: the time cost of the electric vehicle per unit time of travel;

c_w: the penalty cost of unit waiting time for waiting at the boarding point of passengers due to delay of the electric vehicle;

c_e: energy cost per unit of electricity;

c_q: the unit queuing time cost of the electric vehicle in the charging pile queue;

c_f: cost of purchase and use of each electric vehicle；

c_r: the construction and use cost of each charging pile;

t_ij: the travel time from node i e N to node j e N;

the time when the electric vehicle V belongs to the V and actually reaches the node i belongs to the N;

the charging time of the electric vehicle V belongs to V when the charging pile i belongs to C;

the time when the electric vehicle V belongs to V and starts from a passenger boarding point i belongs to P;

a_i: the time when the passenger arrives at the boarding point i belongs to the P;

e: the maximum electric quantity of the electric vehicle;

the electric quantity when the electric vehicle V belongs to the node i belongs to the node N:

a variable of 0 to 1, if the electric vehicle V belongs to the V and directly reaches the node j belongs to the N from the node i belongs to the N, and the variable is equal to 1; otherwise, it equals 0.

r_j: a variable of 0-1 is obtained, and if a charging pile candidate point j belongs to C, a charging pile is installed and is equal to 1; otherwise, it equals 0.

In order to reduce the running time cost of the electric vehicle,

for electric vehicles at charging stationsThe cost of the queuing time of (a),

to penalize the cost for the waiting time of passengers at the boarding point due to the delay of the electric vehicle,

the cost of the charging capacity of the electric vehicle,

for the cost of use of vehicles operating in the road network,

the use cost of the charging pile in the road network is reduced.

1.1.3 sharing passenger-receiving and delivery travel constraint of electric vehicle

And generating a passenger receiving and sending path of the shared electric vehicle by the following constraints, and enabling the vehicle to complete the passenger receiving and sending tasks of each stage according to the path rule shown in the figure 1: at each stage, the electric vehicle v departs from the departure point O, goes to the passenger boarding point P to pick up the passenger, and transports it to the corresponding passenger alighting point D. If the v electric quantity of the electric vehicle is not enough to complete the next task of receiving and sending passengers, the electric vehicle goes to the charging pile C for charging, then goes to other passengers to get on the bus at the bus point P for receiving passengers, and so on. When the electric vehicle v finishes all the passenger receiving and sending demands in the stage, the electric vehicle v stays at the passenger getting-off point D, and the electric vehicle v is guided to return to the virtual terminal T to indicate that the electric vehicle v finishes the passenger receiving and sending tasks in the stage. The passenger getting-off point D where the electric vehicle v stays at the end of the stage is the starting point O of the next stage.

Wherein M represents a sufficient numberA large number. Theta^vThe variable is 0-1, if the electric vehicle V belongs to V to execute the task of receiving and sending the passenger, the variable is equal to 1; otherwise, it equals 0.

The expressions (2) to (4) show that the electric vehicle which starts from the passenger getting-on point i belongs to P can only go to the passenger getting-off point i + delta belonging to D corresponding to the getting-on point i; the electric vehicle which departs from the passenger getting-off point i e D can only go to a charging pile, an electric vehicle returning point or other passenger getting-on points. The above points can be accessed by only one electric vehicle.

The formula (5) shows that the electric vehicle which departs from the charging pile i belongs to C and then can only go to the passenger getting-on point j belongs to P or the electric vehicle returning point j belongs to T. The constraint prevents an electric vehicle departing from charging post i e C from traveling to electric vehicle departure point j e O, passenger disembarkation point j e D, and other charging posts j e C.

The expression (6) shows that the electric vehicle which departs from the electric vehicle departure point i belongs to O can only go to the passenger boarding point j belongs to P or the charging pile j belongs to C. The constraint prevents the electric vehicle which starts from the electric vehicle departure point i belongs to O from moving to the other electric vehicle departure points j belongs to O, the passenger getting-off point j belongs to D and the electric vehicle return point j belongs to T.

The formula (7) shows that the electric vehicle v from the electric vehicle departure point i e O can only go to the getting-on point j e P or the charging pile j e C of one passenger at most. The constraint ensures that the electric vehicles V with the same number can only start from one electric vehicle starting point O, and the number of the vehicles thrown into the road network is more than 0 and less than or equal to V_max. When a new stage begins, if the electric quantity of the electric vehicle is not enough to directly execute a new pick-up task, the electric vehicle can be charged by the formula (6) -formula (7) from the departure point i epsilon O to the charging pile j epsilon C.

Equation (8) indicates that only one electric vehicle v can start from each electric vehicle starting point O. Because each stage needs to update the information of the position, the state and the like of the vehicle, the formula (8) enables the serial number of the departure point O of each electric vehicle to correspond to a different electric vehicle and become a unique index of the electric vehicle, and when the information of the geospatial node, the electric quantity, the departure time and the like corresponding to the same serial number of the departure point O of the electric vehicle is updated, the information update of the electric vehicle corresponding to the serial number can be completed.

Equations (7) to (8) allow the model to determine the optimal fleet size of electric vehicles launched into the road network by controlling the number of electric vehicle departure points O. If a plurality of vehicles are located on a node on the same geographic space, a plurality of different numbers can be assigned to the node to serve as the starting points of different electric vehicles, so that one starting point of the electric vehicle is ensured to correspond to one electric vehicle. The number of electric vehicles performing the pick-up task at each stage is different, and the electric vehicle v may be performing the pick-up task at a certain stage but may be in a standby state without being assigned a task at another stage.

The expression (9) indicates that the electric vehicle does not go to any other point to perform tasks after reaching the electric vehicle return point T.

Equation (10) indicates that if an electric vehicle arrives at the passenger getting-on point P, the passenger getting-off point D, or the charging post C, the vehicle must also leave from the passenger getting-on point P, the passenger getting-off point D, or the charging post C.

The formula (11) and the formula (14) ensure that the electric vehicle can start from the electric vehicle departure point O and finally return to the electric vehicle return point T, and redundant circulation of the electric vehicle path is avoided. If the electric vehicle v performs a task of transporting a passenger from the getting-on point i to the getting-off point i + δ, the electric vehicle is always started from the departure point O (θ)^v1). If the electric vehicle v does not perform any pick-up task, the electric vehicle will stop at the departure point O (theta)^v＝0)。

1.1.4 travel time calculation

Calculating the travel time of the automatic driving electric automobile at each node by the following constraint:

in the formula:

g: charging efficiency of the charging pile, namely, the charging amount per unit time.

v _ start: and tracking and recording the departure time of all electric vehicles before each stage to leave the departure point, and initializing v _ start of all electric vehicles to be 0.

v₀Charge: recording the number of the electric vehicle charged each time in each stage, and initializing the electric vehicle as an empty set;

s₀: record each stageInner electric vehicle v₀Initializing the charging to be an empty set when the charging pile starts to charge;

c₀: record electric vehicle v in each stage₀A charging pile number corresponding to charging is initialized to be an empty set;

y₀charge: record electric vehicle v in each stage₀And initializing the electric quantity when the charge reaches the charging pile to be an empty set.

Equation (15) specifies the time for the electric vehicle to depart from the electric vehicle starting point O at each stage.

The expression (16) to the expression (17) indicates the time from the electric vehicle departure point i e O or the passenger alighting point i e D to the next node j e N for the electric vehicle v

Equal to the time of the electric vehicle v reaching i

Plus the travel time t between i and j_ij。

Equation (18) represents the time for the electric vehicle v to depart from the passenger boarding point i ∈ P

Must be later than the time a when the passenger arrives at the boarding point i ∈ P_i。

Equation (19) represents the time for the electric vehicle v to depart from the passenger boarding point i ∈ P

Must be later than the time when the electric vehicle v arrives at the passenger boarding point i e P

Equation (20) to equation (21) represents the time from the passenger getting-on point i e P to the next node j e N for the electric vehicle

Equal to the electric vehicle v getting on the passengerTime of departure of vehicle point i belonging to P

Plus the travel time t between i, j_ij。

The time that the passenger waits for the electric vehicle v to departure at the boarding point i epsilon P can be calculated to be equal to the time that the passenger waits for the electric vehicle v to departure from the formula (18) and the formula (21)

The formula (22) represents the time when the electric vehicle v starts to be charged in the charging pile i epsilon C

The time that the electric vehicle v arrives at the charging pile i belongs to C must be later

Equation (23) to equation (24) represent the time for the electric vehicle v from the charging pile i e C to reach the next node j e N

Equal to the time that the electric vehicle v starts to charge in the charging pile i belongs to C

Time of charging electric vehicle in charging pile i belonging to C

And a travel time t between i and j_ijAnd (4) the sum.

The formula (25) shows that if the new stage begins, the electric vehicle still has the previous stage to be charged

Charging, namely the electric vehicle v in the new stage is charging pile

The charging must be started later than during the preceding phasesAnd (5) ending the charging time of the electric vehicle charged by the pile.

1.1.5 electric vehicle queue waiting time calculation

Calculating the queuing waiting time of the electric vehicle at each charging station by the following constraints:

in the formula:

is a variable from 0 to 1. If the electric vehicle u belongs to V and reaches the charging station n belongs to V earlier than the electric vehicle V belongs to VU, equal to 1; otherwise, equal to 0;

is a variable from 0 to 1. If the electric vehicle u belongs to the V and reaches the charging pile i belongs to the C earlier than the electric vehicle V belongs to the V, and the value is equal to 1; otherwise, it equals 0.

Equation (26) to equation (29) indicate that if u car arrives at charging pile i ∈ C earlier than v car, the time when v car starts charging at the charging pile must be later than the time when u car finishes charging at the charging pile

On the contrary, if u car arrives at the charging pile i belonging to C later than v car, the time for v car to start charging must be earlier than the time for u car to finish charging

Constraints (30) - (33) prevent the situation as shown in fig. 2 from occurring. Equations (30) - (33) show that if u car arrives at the charging station earlier than v car

Then U car charging pile j belongs to C at charging station n belongs to U_nTime of starting charging

Charging pile i belongs to C earlier than v vehicle at charging station n belongs to U_nTime to start charging

On the contrary, if the u vehicle arrives at the charging station later than the v vehicle

Then U car charging pile j belongs to C at charging station n belongs to U_nTime to start charging

Charging pile i belongs to C at the charging station later than v_nTime to start charging

According to the formula (26) to the formula (33), it can be obtained that the queuing time of the electric vehicle v at the charging station is

1.1.6 electric vehicle electric quantity calculation

The electricity consumption and the charging condition of the shared electric vehicle during the operation process are described by the following constraints:

in the formula:

v _ energy: and tracking and recording the electric quantity of all electric vehicles at the starting point before each stage, and initializing v _ energy of all electric vehicles.

d_ij: the shortest distance between node i ∈ N and node j ∈ N.

h: the electric power consumption rate of the electric vehicle is the electric power consumed by the electric vehicle per unit time.

The expression (34) to the expression (35) indicates the amount of electricity that reaches the next node j ∈ N for the electric vehicle v that departs from the departure point i ∈ O, the passenger boarding point i ∈ P, or the passenger disembarking point i ∈ D

Electric quantity equal to point i

The amount of power consumed by the electric vehicle traveling between ij is subtracted.

Equation (36) indicates that the electric vehicle will be fully charged in charging pile i ∈ C. Because the objective function formula (1) minimizes the electric quantity cost of the electric vehicle for charging in the charging pile, the formula (36) and the objective function ensure that the electric vehicle v which starts from the charging pile i belongs to C and reaches the electric quantity of the next node j belongs to N

Equal to the maximum amount of electric power E of the electric vehicle minus the amount of electric power consumed by the electric vehicle traveling between ij.

And the formula (37) shows that the electric quantity of the electric vehicle v when reaching the destination i e E D of each passenger is enough to support the electric vehicle v to continue to charge any charging pile j E C. Considering that the electric vehicle at each stage will stay at the passenger getting-off point of the last passenger receiving task after finishing all the passenger receiving tasks, and the point is the electric vehicle departure point of the next stage, the electric vehicle at the point is ensured to have enough electric quantity to support the vehicle to continue to execute tasks at the next stage.

Equation (38) represents the electric energy of the electric vehicle at the start point O at each stage.

1.1.7 charging station selection decision

The charging condition of the shared electric vehicle in the charging pile is determined by the following constraints:

and the formula (39) is used for judging whether the electric vehicle goes to a charging pile j epsilon C for charging.

1.2 linking and updating of two-stage information

And after the optimization of one stage is finished, the vehicle is rolled to the next stage for continuous optimization, and in the process, the information such as vehicle path scheduling and charging decision obtained by the optimization of the previous stage needs to be updated to the next stage. And the information such as the position and the electric quantity of the vehicle after the operation of the previous stage is finished is used as an initial value for solving the next stage. As shown in fig. 3, in the previous stage, a virtual return point T is set to record the position number and arrival time of the passenger getting-off point i e D where the electric vehicle put into operation at the stage stays after completing the task of receiving and sending the passenger

And amount of electricity

Setting a virtual departure point O in the next stage, numbering the positions of the passenger alighting points i belonging to D and the arrival time of the electric vehicle staying after completing the task of receiving and sending the passengers in the previous stage

And amount of electricity

The information is passed to the virtual departure point O.

2 solving genetic algorithm of charging station layout scheme

2.1 Individual coding

Each individual is randomly coded using a binary coding rule, with 0 indicating no charging stations (stakes) are built and 1 indicating charging stations (stakes) are built. Variable Z indicating whether or not to establish a charging station_nAnd a variable r indicating whether or not the charging pile is installed_jWill be determined by the individual code values generated. The binary encoding rule enables the generated individuals to satisfy the following two problems while satisfying the equations (40) to (41): 1) at which nodes charging stations are arranged; 2) blockHow many charging piles should be installed to the node of the fixed construction charging station.

In the formula:

Z_nis a variable from 0 to 1. If a charging station is established at n places, the charging station is equal to 1; otherwise, the value is equal to 0,

formula (40) shows that charging piles are installed in the place n only when charging stations are built at the place n epsilon U.

The formula (41) shows that the number of the charging piles built at the charging station n does not exceed the set maximum value

For example, assuming there are 3 charging station alternate points in total, [1,0,1]Representing that charging stations are built at No. 1 and No. 3 nodes, no charging station is built at No. 2 node, Z_nThe values of (a) are 1,0,1, respectively. If a certain node is not provided with a charging station, the node is not provided with a charging pile, so that the power station is [1, 1]],[0,0,0],[1,0,1]]Show that 3 are installed at No. 1 node and are filled electric pile, No. 2 nodes are not built a station and are built a stake, and 2 are installed at No. 3 nodes and are filled electric pile, r_jAre 1,1,1,0,0,0,1,

the individual codes also need to satisfy the following two constraints of formula (42) and formula (43), so that the total number of charging stations and charging piles is limited, and the algorithm is prevented from being borderless when the optimal solution is searched.

Equation (42) indicates that the number of charging stations established at all charging station candidate points cannot exceed C_max。

Equation (43) indicates that the number of charging piles constructed at all charging stations cannot exceed Q_max。

All solution individuals violating the constraint are assigned a maximum value, and by doing so, the infeasible solution is screened out in the subsequent selection, hybridization and mutation operations.

2.2 iterating using evolutionary operations

And carrying out genetic evolution operations such as selection, hybridization, mutation and the like on each generation of population, and forming a new generation of population after reinsertion. And performing iterative evolution until the minimum value of the individual fitness in the population reaches convergence. The charging station layout scheme represented by the individual with the minimum fitness value is the optimal layout.

Summary of large-scale road network solution method

Summarizing the solution method for optimally laying intelligent network connection shared electric vehicle charging stations in a large-scale network, which comprises the following steps:

wherein:

v _ origin: tracking and recording the starting point numbers of all the electric vehicles before each stage;

v _ terminal: recording the final return points of all the electric vehicles after all the electric vehicles execute all the passenger travel requirement tasks, wherein the final return points are equal to the departure points of all the electric vehicles provided in the lead-in road network;

v₀: recording all vehicles operating in each phase;

v₀origin: recording the starting points of all vehicles running in each stage;

v₀α destin: recording the end points of all vehicles operating in each phase;

y₀_ destin: recording the electric quantity of all running vehicles reaching the terminal point in each stage;

tao₀α destin: recording the time of all running vehicles reaching the terminal point in each stage;

c₀used: and recording the charging piles charged by all running vehicles in each stage.

The embodiment is as follows:

the Chicago road network with 933 nodes, 2967 road segments and 300 passenger travel needs is solved. The number, longitude and latitude coordinates, road section length and passenger travel demand OD peer information of each node of the road network refer to data provided by a Transportation Networks column on a Github website.

And throwing 20 intelligent networked shared electric vehicles in a road network, selecting 20 nodes as a starting point/returning point of the electric vehicle, and selecting 10 nodes as alternate points of an electric vehicle charging station. The numbers and coordinates of the selected alternative points of the charging station and the departure/return points of the electric vehicle are shown in tables 1 and 2, respectively. The topology of road network and the layout of the electric vehicle starting point/returning point and the charging station alternative point are shown in FIG. 4, wherein the triangle represents the electric vehicle starting point/returning point, the dot represents the charging station alternative point, and the number beside the dot represents the charging station alternative point corresponding to the tableNumber in 1. Defining the maximum number of stations C that can be set up in the road network_max8, the electric pile quantity R that fills that can install at most of every charging station_max3, the number of charging piles installed at all charging stations does not exceed Q_max＝30。

TABLE 1 number and coordinates of alternate points of Chicago road network electric vehicle charging station

TABLE 2 number and coordinates of departure/return points of Chicago road network electric vehicle

Setting the duration of each stage to be 60 time units, and enabling 300 travel demands to be according to the time a when the passengers arrive at the boarding point_iDivided into various stages. Setting the speed of all electric vehicles to 5, and setting the time cost c per unit travel time_tEnergy cost per unit of electricity c1 _e1, cost per electric vehicle used c_fPenalty cost per delay time c of 100_pCost per queuing time c2_qPenalty cost c in waiting time of getting-on point unit 2_w2, construction and installation cost c for each charging station_s80, cost of use c per charging pile_rThe charging efficiency g is 10, the power consumption efficiency h is 10, and the maximum electric quantity E of the electric vehicle battery is 500. Setting 10 individuals in each generation of population, wherein the iterative evolution times of the population is 50, and the probability of crossover and variation is 90%. And (3) building a genetic algorithm framework by using a DEAP evolution algorithm tool kit based on python API to perform population iterative evolution, and solving the minimum value of the evaluation function by using a Dual Simplex method (Dual Simplex) solver of Gurobi operational optimization software.

And solving a result, along with the increase of the iteration times of the population, the minimum value, the average value, the maximum value and the standard deviation of the fitness of all individuals of each generation of the population are shown in a table 3, and the convergence condition of the fitness value of each individual is shown in a figure 5. It can be found that when the 9 th generation is reached, the minimum value of the population individual fitness reaches convergence, and the convergent individual is coded as [ [0,0,0], [1,1,1], [1,0,1], [0,0,0], [0,0,0], [1,0,1], [1,0,1], [1,0,1], [0,0,0] ], i.e. a total of 4 charging stations are established: the charging station of serial number 569 installs 3 and fills electric pile, and the charging station of

serial number

892, 786, 584 all installs 2 and fills electric pile. The optimal layout of the charging stations is shown in fig. 6, the dots represent the charging stations constructed in the road network, and the numbers beside the dots represent the charging stations corresponding to the numbers in table 1. The total cost of this charging station layout is also 127389.2.

TABLE 3 variation of fitness value of population individuals with iteration number

It is to be understood that the present invention has been described with reference to certain embodiments and that various changes in form and details may be made therein by those skilled in the art without departing from the spirit and scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from the essential scope thereof. Therefore, it is intended that the invention not be limited to the particular embodiment disclosed, but that the invention will include all embodiments falling within the scope of the appended claims.

Claims

1. An optimal layout solving method for intelligent network connection sharing electric vehicle charging stations in a large-scale network is characterized by comprising the following steps:

c2: solving the established model by combining a time window rolling staged dynamic programming method and a genetic algorithm; the upper layer generates an initial population containing a plurality of individuals by using a genetic algorithm, wherein each individual represents a charging station layout scheme; according to an individual fitness evaluation value provided by a lower-layer problem, carrying out genetic evolution operations of selection, hybridization and variation on each generation of population, and forming a new generation of population after reinsertion; performing iterative evolution until the minimum value of the individual fitness in the population converges; the charging station layout scheme represented by the individual with the minimum fitness value is the optimal layout;

the lower layer solves a mixed integer linear programming Model (MILP) after an individual layout scheme is given by using a time window rolling staged dynamic programming method, so that the fitness of each individual is obtained; under each individual scheme, dividing a time period covered by a Model (MILP) into a plurality of small stages, and simultaneously dividing large-scale passenger trip data into a plurality of small data according to each time stage; in each stage, solving a mixed integer linear programming Model (MILP) with reduced scale according to the current vehicle state of a road network and the travel demand information of passengers in the stage, and after the solution of each stage is finished, correspondingly updating vehicle path scheduling and charging decisions and state information of all vehicles to serve as an initial value of the solution of the next stage; in the same way, the original large-scale problem is split in the time dimension by a method of continuously rolling in the time stage, and the high dimension is reduced to the low dimension; calculating and adding the obtained objective function values of all stages to obtain a fitness value of each population individual, wherein the fitness value is the total cost of the charging station layout strategy represented by the individual;

c3: and summarizing a large-scale road network solving method.

2. The optimized layout solution method for the intelligent networked shared electric vehicle charging stations in the large-scale road network according to claim 1, wherein the mixed integer linear programming Model (MILP) comprises building of a traffic network module, an objective function module, an electric vehicle passenger receiving and sending path module, a travel time module, an electric vehicle queuing and waiting time module, an electric vehicle electric quantity module and a charging station module;

c1.1: building a traffic network module: the optional departure point of the electric vehicle is represented by the parameter O, P represents the set of boarding points for all passengers, D represents the set of disembarking points for all passengers, U represents the set of backup point numbers for all charging stations,

the maximum number of charging piles which can be installed at the alternative points of the charging station n is represented; v_maxIndicating the number of electric vehicles which are allowed to be released in the network at most; c_maxIndicating that a limit on the number of charging stations allowed to be established in the road network is set; c_nRepresenting a set of all alternative charging piles of the alternative points of the charging station number n; q_maxShowing the quantity limit of the charging piles put in a road network; c represents all charging station charging pile alternate point sets; t represents a return point of the electric vehicle, which is a virtual terminal point, and the distance from the electric vehicle to each other node is 0; n represents a set of all nodes; v represents the set of all the alternative electric vehicles, and a digital model of a physical traffic network is established;

c1.2: and (3) building an objective function module: by a parameter c_t、c_w、c_e、c_q、c_f、c_r、t_ij、

a_i、E、

r_jCalculating a minimum value of the sum of total costs of the respective stages; namely the running time cost of the electric vehicle, the use cost of the automatic driving electric vehicle and the charging pile, and the electricityThe minimum value of the sum of the charging cost of the bullet train at the charging station, the waiting time cost of the electric vehicle at the charging station and the penalty cost generated by the node which cannot be reached on time is c_tRepresenting the time cost of the electric vehicle running for unit time; c. C_wThe penalty cost of unit waiting time for waiting at the boarding point of passengers caused by delay of the electric vehicle is represented; c. C_eRepresents the cost of energy consumption per unit of electricity; c. C_qThe unit queuing time cost of the electric vehicle in the charging pile queuing is shown; c. C_fRepresents the purchase and use cost of each electric vehicle; c. C_rThe construction and use cost of each charging pile is represented; t is t_ijRepresenting the travel time from node i e N to node j e N;

representing the time when the electric vehicle V belongs to V and actually reaches a node i belongs to N;

representing a variable of 0-1, if the electric vehicle V belongs to V and directly reaches a node j belongs to N from a node i belongs to N, and the value is equal to 1; otherwise, equal to 0; r is_jRepresenting a variable of 0-1, and if a charging pile is installed at a charging pile candidate point j epsilon C, the variable is equal to 1; otherwise, equal to 0;

c1.3: the electric vehicle receives and sends the building of the passenger path module: and (3) generating a passenger receiving and sending path of the shared electric vehicle through formula calculation constraint, and enabling the vehicle to complete the passenger receiving and sending tasks of each stage according to a specified path rule: at each stage, the electric vehicle v starts from a departure point O, goes to a passenger boarding point P to pick up passengers, and conveys the passengers to a corresponding passenger alighting point D; if the electric quantity v of the electric vehicle is not enough to complete the next task of receiving and sending the passengers, the electric vehicle goes to a charging pile C for charging, then goes to other passengers to get on the bus at a bus point P for receiving the passengers, and so on; when the electric vehicle v finishes all the passenger receiving and sending demands in the stage, the electric vehicle v stays at a passenger getting-off point D, and the electric vehicle v is guided to return to a virtual terminal T to represent that the electric vehicle v finishes the passenger receiving and sending task in the stage; the passenger getting-off point D where the electric vehicle v stays at the last stage is the starting point O of the next stage;

c1.4: building a travel time module: by the parameters g, v _ start, v₀_charged、s₀、c₀、y₀Charging, calculating the travel time of the automatic driving electric automobile at each node; wherein g represents the charging efficiency of the charging pile, i.e., the charging amount per unit time; v _ start represents the departure time of all electric vehicles leaving the departure point before tracking and recording each stage, and v _ start of all electric vehicles is initialized to be 0; v. of₀The charge represents the number of the electric vehicle which is charged each time in each stage and is initialized to be an empty set; s₀Showing and recording electric vehicle v in each stage₀Initializing the charging pile to be an empty set when the charging pile starts to charge; c. C₀Showing and recording electric vehicle v in each stage₀A charging pile number corresponding to charging is initialized to be an empty set; y is₀"charged" means that the electric vehicle v in each stage is recorded₀Initializing the electric quantity when the charge reaches the charging pile to be an empty set;

c1.5: the electric vehicle queuing waiting time module is built: by the parameters and parameters described in C1.1-C1.4

And

Is a variable from 0 to 1;

c1.6: electric vehicle electric quantity module is built: by the parameters described under C1.1-C1.4 and the parameters v _ energy, d_ijAnd h, describing the electricity consumption and the charging condition of the shared electric vehicle in the operation process, wherein v _ energy represents the electricity quantity of all the electric vehicles at the starting point before each stage of tracking and recording, and v _ energy of all the electric vehicles is initialized to E and d_ijThe shortest distance between the node i belonging to N and the node j belonging to N is represented, and h represents the electric quantity consumption rate of the electric vehicle, namely the electric quantity consumed by the electric vehicle in unit time;

c1.7: building a charging station module: passing parameters

3. The optimized layout solution method for intelligent networked shared electric vehicle charging stations in large-scale road network according to claim 1, wherein said step C2 comprises the following steps:

c2.1: individual coding;

c2.2: the iteration is performed using an evolutionary operation.

4. The optimized layout solution method for the intelligent networked shared electric vehicle charging stations in the large-scale road network according to claim 3, wherein the step C2.1 comprises the following steps: each individual is randomly coded by using a binary coding rule, 0 represents that no charging station or charging pile is built, 1 represents that a charging station or charging pile is built, and a variable Z represents whether a charging station is built or not_nAnd a variable r indicating whether the charging pile is installed or not_jWill be determined by the individual code values generated.

5. The optimal layout solution method for the intelligent network interconnection sharing electric vehicle charging stations in the large-scale network according to claim 3, wherein the step C2.2 is to perform genetic evolution operations of selection, hybridization and variation on each generation of population, form a new generation of population after reinsertion, and perform iterative evolution in such a way until the minimum value of the individual fitness in the population converges, and the layout scheme of the charging station represented by the individual with the minimum fitness value is the optimal layout.

6. The optimized layout solution method for the intelligent network interconnection sharing electric vehicle charging stations in the large-scale network according to claim 1, wherein the step 3 comprises the following specific steps:

c3.1: initializing;

c3.2: individual coding;

c3.4: performing population iteration by using a genetic algorithm;

c3.5: and obtaining an optimal charging station layout scheme.

7. The optimized layout solution method for the intelligent network interconnection sharing electric vehicle charging stations in the large-scale network according to claim 6, wherein the C3.1 comprises the following specific steps:

c 3.1.1: introduction and study of road network:

obtaining the shortest paths between nodes such as m-hour passenger travel demand OD pairs and departure time thereof, electric vehicle departure points or return points, all charging station alternative points and the like in a road network and each node, and calculating to obtain the layout schemes of all charging stations;

c3.1.2: a splitting stage:

c3.1.3: initializing a layout scheme of a charging station:

and randomly generating an initial charging station layout scheme.

8. The optimized layout solution method for the intelligent networking sharing electric vehicle charging stations in the large-scale network as claimed in claim 6, wherein the C3.2 comprises the following specific steps:

c3.2.1: coding of charging station layout:

encoding the initial charging station layout scheme;

c3.2.2: the code that electric pile was laid:

9. The optimized layout solution method for the intelligent networked shared electric vehicle charging stations in the large-scale road network according to claim 6, wherein the C3.3 adopts the established mixed integer linear programming model with the total cost minimized as a target as an evaluation function of each individual fitness, and comprises the following specific steps:

c3.3.1: initializing a relevant set; initializing the set v _ origin, v _ start, v _ energy, v _ terminal, v _ origin₀_charged，s₀，c₀，y₀Charging; establishing an aggregate cost to record the total cost of the electric vehicle for completing all passenger pick-up and delivery requirement tasks under the charging station layout scheme u, and initializing the cost to be 0;

c3.3.2: calculating a target function value of the electric vehicle for receiving passengers in stages;

each stage in the split s stages is k, and the following steps are executed in a circulating mode:

c3.3.2.2: optimizing the implementation stage; solving a mixed integer linear programming Model (MILP) of the stage to obtain a vehicle passenger receiving and sending path strategy, an electric vehicle scale, a charging strategy and an objective function value k _ cost which are optimal in the stage k, and updating the cost as cost + k _ cost;

c3.3.2.3: acquiring information of the vehicle running at the stage k; acquiring a vehicle running in the stage k, a starting point of the running vehicle, a terminal point of the running vehicle, electric quantity of the running vehicle reaching the terminal point, time of the running vehicle reaching the terminal point and a charging pile for charging the running vehicle, and respectively importing corresponding information into a set v₀,v₀_origin,v₀_destin,y₀_destin,tao₀_destin,c₀Used;

c3.3.2.6: updating the starting time of the electric vehicle; the departure time v _ start of all vehicles in the next stage k +1 is updated to t (k +1), and the time tao when the running vehicle reaches the end point is further updated₀Comparing _destinwith t · (k + 1); if tao₀α destin > t · (k + 1): operating the vehicle v in phase k₀Time to endpoint tao₀Update to next phase k +1 operating vehicle v₀In the departure time v _ start; otherwise, v _ start remains unchanged;

c3.3.2.7: removing the records of the electric vehicle which finishes charging in the stage k; charging pile c of electric vehicle₀Used end time of charging s₀+(E-y₀) Comparison of/g with the start time t of stage k +1 (k + 1): if s is₀+(E-y₀) T (k + 1): delete electric vehicle at v₀_charged，s₀，c₀，y₀All records in _ charged; otherwise, these records are kept;

c3.3.3.1: updating the traffic network; introducing the updated v _ origin set, constructing each node set in sets O, C, T and N according to the traffic network construction rule, and updating the shortest path of each node of the constructed traffic network;

wherein:

v _ terminal represents a return point which is finally returned after all electric vehicles execute all passenger travel requirement tasks and is recorded, and the return point is equal to the starting point of all the electric vehicles provided in the lead-in road network;

v₀representing all vehicles that are logged as operating in each phase;

10. The optimized layout solution method for the intelligent networking sharing electric vehicle charging stations in the large-scale network according to claim 6, wherein the iteration times gen of the C3.4 set population are iterated from the second generation, and the method comprises the following specific steps:

c3.4.3: carrying out variant evolution; carrying out mutation on the crossed individuals in a disorder mutation mode; recalculating the fitness of the changed individual;

c3.4.4: carrying out reinsertion: selecting w individuals with the best fitness in the breeding offspring, inserting the w individuals into the parent, removing the individuals with the same codes, and taking the obtained result as the parent population for next iteration; the C3.5 specific steps are as follows: if the minimum value of the individual fitness in the population converges after iteration, the individual code with the minimum fitness in the last generation of population is considered as the optimal layout scheme, the layout scheme is output, and if the minimum value of the individual fitness in the population does not reach convergence, C3.5 is repeated until the convergence is reached.