Expressway service area charging pile configuration method and system integrating power allocation and multi-agent system

Abstract

The invention discloses an expressway service area charging pile configuration method and system integrating power allocation and a multi-agent system. According to the invention, the Agent-based Model (ABM) is used to simulate and describe the driving and charging behaviors of the new energy vehicle on the highway, and the multi-objective optimization algorithm is used to optimize the charging pile arrangement scheme. On the basis of the multi-agent system, the charging demand end of interaction of the vehicle, the road and the charging pile is considered, and the power distribution system (charging supply end) of the power grid, the service area and the charging pile is established, so that power distribution is carried out on the service vehicle under the maximum load power of the power grid; a set of distribution rules is established under the consideration of fair and overall effects, and the configuration of the charging piles in the highway service area is optimized in combination with an existing optimization algorithm, so that the purposes of meeting the charging requirements of users and optimizing resource configuration are achieved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation technology and relates to a method and system for configuring charging piles in highway service areas that integrates power allocation and multi-agent systems. Background Technology

[0002] The development and promotion of electric vehicles is an important measure for the green transformation of the transportation industry, and the widespread use of electric vehicles is inseparable from the supporting development and improvement of charging infrastructure. In existing research, the optimization of charging pile configuration in highway service areas often adopts a combination of simulation and optimization algorithms. In terms of simulation, existing studies often only consider the relationship between vehicles, roads, and charging piles. Expanding the number of charging piles not only incurs huge financial costs, but also affects the power supply system. If too many charging piles are added, the excessive load on the power grid will reduce service quality and pose safety hazards. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention proposes a method and system for configuring charging piles in highway service areas that integrates power allocation and multi-agent systems.

[0004] The method for configuring charging piles in highway service areas using fusion power allocation and multi-agent systems in this invention includes the following steps:

[0005] S1. Construct a multi-agent simulation environment consisting of a highway agent, a vehicle agent, and a power dispatching system agent, where:

[0006] The highway agent is used to provide a static road network topology including nodes, road segments and service areas, and dynamically updates the road segment speed using the BPR function based on real-time traffic flow.

[0007] The vehicle agent is used to generate electric vehicle entities based on the OD matrix and update the location and SoC in real time during driving. When the vehicle enters the service area, it triggers a charging decision based on the mandatory charging threshold, the optional charging threshold and the probabilistic rules.

[0008] The power allocation system Agent is used to dynamically allocate charging power requests in each service area according to the two-level proportional allocation principle of "grid-service area-charging pile" under the constraint of the maximum power available by the power grid Pmax.

[0009] S2. To minimize the weighted sum of average user queuing time and charging pile construction and operation costs, an optimization objective function including a queuing time penalty term is established, and the particle swarm optimization algorithm is used to iteratively optimize the number and type ratio of charging piles in each service area.

[0010] S3. Output a charging pile configuration scheme that satisfies the upper and lower limits of the number of charging piles in the service area and minimizes the objective function.

[0011] A highway service area charging pile configuration system according to the present invention includes:

[0012] The multi-agent simulation module is used to build and run an interactive environment consisting of a highway agent, a vehicle agent, and a power dispatching system agent, wherein:

[0013] The highway agent is configured to dynamically update the road segment speed by calling the BPR function based on real-time traffic flow, and to provide route and speed information to the vehicle agent.

[0014] The vehicle agent is configured to update its location and battery status during driving, and to trigger a charging decision based on a mandatory charging threshold, an optional charging threshold, and probabilistic rules when entering a service area.

[0015] The power allocation system Agent is configured to dynamically allocate charging power requests at each level according to the two-level proportional allocation principle of "grid-service area-charging pile" under the constraint of the maximum available power Pmax of the power grid;

[0016] The optimization calculation module is used to run the particle swarm optimization algorithm with the goal of minimizing the weighted sum of the average user queuing time and the construction and operation cost of charging piles. It iteratively optimizes the number and type ratio of charging piles in each service area and outputs the optimal configuration scheme.

[0017] The data interface module is used to input traffic demand data, power grid parameters and charging pile parameters into the multi-agent simulation module, and to write the configuration scheme obtained by the optimization calculation module into an external database or scheduling system.

[0018] The beneficial effects of this invention are as follows: By establishing a multi-agent simulation system, this invention not only considers the charging demand side of the interaction between vehicles, road networks, and charging piles, but also the charging supply side of the power grid, service areas, and charging piles. Through optimization algorithms, it optimizes the configuration of charging piles in various service areas under the limited load power of the power grid, thereby achieving the goal of meeting users' charging needs and promoting the improvement of electric vehicle supporting facilities. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a highway.

[0020] Figure 2 This is a topology diagram of the highway network.

[0021] Figure 3 A topology diagram of the road network and service areas;

[0022] Figure 4 Maps showing vehicle origin-destination (OD) and 24-hour traffic demand distribution;

[0023] Figure 5 A comparison chart of charging pile configurations before and after optimization;

[0024] Figure 6 The graph shows the convergence curve of the algorithm.

[0025] Figure 7 This is a comparison chart of average waiting times;

[0026] Figure 8 This is a comparison chart showing the proportion of people not queuing.

[0027] Figure 9 A satisfaction rate comparison chart;

[0028] Figure 10 A comparison chart of maximum queue lengths;

[0029] Figure 11 A cost comparison chart of charging piles;

[0030] Figure 12 A comparison chart of charging pile utilization rates;

[0031] Figure 13 This is a comparison chart of the overall system performance. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be further described in detail below through embodiments and in conjunction with the accompanying drawings. It should be understood that the specific embodiments described herein are only one preferred embodiment of the present invention and are only used to explain the present invention. They do not limit the scope of protection of the present invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] The basic idea of ​​this application is to simulate and characterize the driving and charging behavior of new energy vehicles on highways using an Agent-based Model (ABM), and to optimize the deployment scheme of charging piles using a multi-objective optimization algorithm. Unlike existing simulation schemes, this application, based on a multi-agent system, not only considers the charging demand side of the interaction between vehicles, roads, and charging piles, but also establishes a power allocation system (charging supply side) connecting the power grid, service areas, and charging piles. Under the maximum load power of the power grid, power allocation is performed on service vehicles. A set of allocation rules is established considering fairness and overall effect, and combined with existing optimization algorithms, the configuration of charging piles in highway service areas is optimized to achieve the goal of meeting user charging needs and optimizing resource allocation.

[0034] Based on the above concept, the specific steps of the method in this application are as follows:

[0035] C1. Highway Agent, Vehicle Agent, Power Distribution System Agent Settings

[0036] C11 Highway Agent Settings

[0037] As a core component of the simulation system, the highway agent provides a digital traffic environment for the vehicle agent to operate in, and undertakes important tasks such as spatial definition, path guidance and traffic condition control.

[0038] C111 Static Road Network Construction

[0039] Static road network construction is the foundation for the functionality of highway agents. The road network is broken down into two basic data classes: Node (i.e., toll booths) and Link (road segment). The Link class records the physical attributes of the road segment in detail, including static parameters such as the starting and ending nodes of the connection, actual length, number of lanes, and legal speed limit.

[0040] At the topology graph integration level, the system incorporates all Node and Link objects into a Network-X directed graph (Di-Graph) structure. In this graph, toll booths exist as vertices, and road segments are represented as directed edges connecting the vertices, with the weight of the edges set to the road segment length.

[0041] C112 Route Planning Service

[0042] The route planning service provides vehicle agents with a defined travel path from the origin to the destination. When a vehicle agent is generated, it sends a route query request to the road segment agent based on the origin and destination. After receiving the request, the road network agent uses the topology of its internal Network-X graph to call Dijkstra's shortest path algorithm, calculates the optimal path with road segment length as the weight, and returns it to the vehicle agent as a sequence of path nodes as its fixed travel route.

[0043] C113 Dynamic Congestion Feedback

[0044] Dynamic congestion feedback is a core dynamic function of the road network agent, enabling the road network to respond to changes in traffic flow. The system records the current number of vehicles on each road segment in real time each day. When a vehicle agent queries the road segment's speed, the road network agent calls the BPR function, taking the road segment's design capacity (determined by the number of lanes and speed limits) and real-time traffic flow as input parameters to calculate the current actual speed.

[0045]

[0046] in, It is the average speed of the road segment. It is the free-flow speed of the road segment. It refers to the traffic flow of the road segment. It refers to the road section's traffic capacity. , It is the hindrance coefficient.

[0047] C12 Vehicle Agent Settings

[0048] The vehicle agent is the core component simulating the driving and charging behavior of vehicles in a highway network. Its design fully considers key factors such as the vehicle's driving path, speed, state of charge (SoC), and service area charging decisions, enabling dynamic simulation of a single vehicle's energy consumption and charging behavior from the entrance node to the exit node.

[0049] C121 vehicle generation

[0050] The vehicle generation process is based on real traffic demand, and its core lies in transforming the preprocessed Origin-Destination (OD) analysis results into concrete vehicle entities. The "Vehicle Generator" module in the system dynamically acquires traffic demand data for the current simulation period through real-time interaction with the OD analyzer.

[0051] C122 vehicle attributes assigned

[0052] At the identification level, the entire lifecycle of a vehicle is tracked through a unique ID; at the trip planning level, the starting point (entry_node) and the ending toll station (exit_node) are clearly defined based on OD requirements to form a complete route guide; at the vehicle type level, the vehicle types are divided into three categories—passenger vehicles, freight vehicles, and special-purpose vehicles—according to my country's current "Classification of Vehicle Types for Toll Roads" (JT2019-02), and are identified in the model through the vehicle type attribute (type).

[0053] The system uses a preset "electric vehicle penetration rate" parameter (ev_penetration_rate) combined with vehicle model-specific settings to determine the electrification of each generated vehicle. For Agents determined to be electric vehicles, their electrical attributes are automatically initialized, including battery capacity matching the vehicle model, energy consumption per 100 kilometers, and initial battery level (initial_soc).

[0054] C123 vehicle location update

[0055] Within a given time period, the vehicle agent and the road network agent dynamically interact to obtain the current average speed of the road segment. Based on the obtained dynamic speed, the distance the vehicle travels per unit time is calculated and added to the vehicle's position_on_link attribute to update the position. When the accumulated value of position_on_link exceeds the length of the current road segment, it is determined that the vehicle has reached the next toll node.

[0056] C124 vehicle power consumption

[0057] During a vehicle agent's operation, the most important metric to monitor is the change in the vehicle's state of charge (SoC). Existing research has shown the following relationship between SoC changes and distance traveled and average speed when a vehicle is traveling on a highway:

[0058]

[0059] in, This refers to the energy consumption (kWh) of an electric vehicle traveling on the road network. Average speed of vehicles traveling on the road segment related.

[0060] In practical implementation, after the vehicle completes its location update every minute, the system uses the newly generated travel distance and average speed of the road segment as the basis for the update. Perform energy consumption calculations.

[0061] C125 Vehicle Charging Decision

[0062] The relationship between SoC (Systemic Cost of Vehicles) changes and distance traveled and average speed when a vehicle is traveling on a highway can be roughly expressed by the following formula:

[0063]

[0064] The formula has been given above. The calculation formula, It is related to the average speed of the vehicle traveling in segment ij. Let n be the battery capacity (kWh) of vehicle n. In the simulation, the above formula can be used to calculate the vehicle's SoC changes in real time.

[0065] The charging decision module simulates the charging choices of electric vehicle drivers at service areas, serving as a crucial link between traffic flow and charging demand. This module is triggered when the vehicle reaches the service area coordinates and, through a multi-level decision-making and planning system, simulates the driver's charging choice logic under different scenarios. Hierarchical decision-making and planning:

[0066] "Forced charging" rule: When the vehicle's current SoC is lower than the preset "forced charging threshold" soc_trigger_threshold, the system determines that the vehicle is at risk of running out of power and will trigger a charging decision with a 100% probability.

[0067] "Random charging" rule: When the SoC is higher than the forced charging threshold but lower than the "optional charging threshold" soc_optional_threshold, the system makes a probabilistic decision based on the optional_charge_probability parameter to simulate the driver's "charge on the way" behavior.

[0068] "Ignore charging" rule: When the SoC is higher than the optional charging threshold, the system determines that the battery is sufficient to support the remaining range, and the vehicle will not choose to charge and continue driving.

[0069] C126 vehicle charging time

[0070] The vehicle's charging time is related to the vehicle's SoC at the service area, the target SoC, and the power of the charging station, and is calculated using the following formula:

[0071]

[0072] in, It is the vehicle target SoC. It is the current SoC. It refers to the vehicle's battery capacity. That is the power of the charging station.

[0073] C13 Power Distribution System Agent Settings

[0074] C131 System Composition and Responsibilities

[0075] In the simulation system, the power allocation system agent consists of three parts: the power grid, the service area, and the charging piles. It plays a crucial role in providing and allocating limited resources. Its core function is as a charging resource manager, with primary responsibilities including physical representation, resource management, and service provision.

[0076] In terms of physical representation, the power distribution system agent stores the geographical location of the service area and its available charging facilities. In terms of resource management, it is responsible for managing its limited internal charging resources (the maximum power that the power grid can provide). The system responds to charging requests from vehicle agents. Service provision involves offering charging services to vehicle agents that have successfully acquired resources; this service process will consume resources for a specific period of simulation time.

[0077] C132 power distribution method

[0078] In resource management, the power allocation system agent achieves multi-level power allocation through the power grid, service areas, and charging piles. Vehicle agents directly request power from charging piles, service areas calculate the total power required by all charging piles and request it from the power grid, which then allocates power to the service area and then to the charging piles. Based on considerations of fairness and overall resource efficiency, the following power allocation rules are established:

[0079] "Grid-Service Area Level": If the total power requested by all service areas is less than the maximum power supplied by the grid. If all service areas are satisfied, then the basic power allocation will be adjusted; otherwise, the basic power allocation will be adjusted accordingly. Power is allocated according to the power request ratio of each service area, with additional power allocated accordingly. The allocation is based on the total number of vehicles in service and queuing in the service area during the current time period, from highest to lowest, with an allocation coefficient of [missing value]. , , , ...

[0080] "Service Area - Charging Pile Level": Allocated according to the proportion of power requested by each charging pile.

[0081] C2. Detailed Simulation Process

[0082] C21 Road Network Construction. The highway network is initialized by decomposing it into nodes (i.e., toll stations) and road segments. Node information includes toll station ID and name, while road segment information includes name, start node, end node, segment proportion (actual length percentage), number of lanes in both directions, and speed limit information. Service area names and locations are also configured to construct a complete road network topology.

[0083] C22 Data Preparation and Processing. The acquired ETC data is processed, invalid data is removed, and valid data including vehicle ID, origin point, and vehicle type is retained. A vehicle OD matrix is ​​generated based on the vehicle's origin point, and OD pairs outside the road network are mapped to the nearest boundary node. Simultaneously, charging facility data (number of fast chargers, superchargers, etc.) for each service area is loaded.

[0084] C23 Simulation Parameter Settings. This section sets parameters for electric vehicles, charging behavior, charging stations, and battery regeneration (BPR). Electric vehicle parameters include penetration rates for different models, battery capacity, and energy consumption. Charging behavior parameters include SoC forced charging threshold, optional charging threshold, and target SoC. Charging station parameters include charging station power and driver charging station selection preferences.

[0085] C24 Vehicle Simulation Flow. Vehicles are generated based on the vehicle OD matrix, including origin, destination, and vehicle type information. Whether a vehicle is an electric vehicle is randomly determined based on penetration rate. If an electric vehicle is identified, an initial SoC is set. The shortest path algorithm is used to obtain the path based on the origin. The vehicle then travels along a road segment, calculating the average speed based on the current segment information using the BPR function. The vehicle's position and battery level are updated every unit of time. The system checks if the vehicle has passed a service area based on its current location and location. If it has, it is determined to be within the service area range, and a charging decision is made. If charging is not selected, the vehicle continues driving; otherwise, it enters the charging service process. If an available charging station is available, the vehicle can directly enjoy the charging service. If no charging station is available, the vehicle must wait in line. Charging time is calculated based on the current vehicle SoC, the target SoC, and the charging station's power. The simulation returns to the road segment driving phase if no charging is needed or if charging is complete. The simulation ends when the vehicle reaches its destination node.

[0086] The C25 simulation has ended. The simulation ends when the set final time is reached.

[0087] C3. Optimization algorithms for finding optimal solutions

[0088] The C31 objective function is as follows: Based on the primary objective of optimizing user queuing time, and considering the limitation of service area size on the number of charging piles, an optimization objective function is constructed to minimize the weighted sum of average queuing time and charging pile construction and operation costs. The specific mathematical model is as follows:

[0089]

[0090]

[0091]

[0092]

[0093] The objective function mainly consists of two parts: average user queuing time and charging station construction and operation costs. These are then normalized. This is the sum of the costs of charging stations. This represents the maximum acceptable cost for charging stations. Average user queuing time The maximum acceptable queuing time for user satisfaction. and That is the weight of the two. As a penalty item, This is a penalty coefficient, applied to service areas where the average queuing time exceeds the maximum acceptable queuing time for users. In practice, factors such as queuing time, maximum queuing length, and queuing ratio can be considered comprehensively.

[0094] In practice, constraints such as the size of the service area and the amount of capital invested will limit the number of charging stations. Therefore, the number of charging stations needs to meet the following constraints:

[0095]

[0096] This represents the number of charging stations in the i-th service area.

[0097] C32 Particle Swarm Optimization Algorithm Design. Particle Swarm Optimization (PSO) is a stochastic search algorithm that simulates swarm behavior, achieving global optimum search through information sharing among particles. Each particle represents a solution, updating its position by flying around the historical best and the global best. PSO performs well in continuous space problems, has fewer parameters, and converges quickly, making it suitable for handling the low-dimensional, integer approximation optimization problem presented in this application. The particle's position is represented as an n-dimensional vector. This corresponds to the number of charging piles in n service areas. Although the particle position is essentially a floating-point number, it is rounded down during the fitness evaluation phase to accommodate the integer number of charging piles.

[0098] The particle's velocity and position are updated using the following formula:

[0099]

[0100]

[0101] in Inertial weights are used to balance exploration and development; These are individual experience factors and group experience factors, respectively. The coefficients are independent and uniformly distributed, which enhances randomness; This represents the optimal position in the particle's history. The position is the global optimum. In each iteration, if the fitness of a particle is better than its historical record or the global optimum, the individual or group optimum will be updated accordingly until the last iteration is completed, and the group optimum will be output as the found optimal solution.

[0102] C4. Evaluation Index Analysis

[0103] C41 Queue Evaluation Indicators. Various queue indicators reflect the service quality level of a service area and also indicate how well users' charging needs are being met.

[0104] C411 Average Waiting Time. Average waiting time is the ratio of total waiting time to the number of vehicles waiting in line. This is the main optimization objective of the objective function, reflecting the overall queuing time for waiting users.

[0105] C412 No-Queue Ratio. The no-queue ratio refers to the proportion of users who need charging services but do not need to queue. This indicator reflects the situation of users enjoying a higher level of service.

[0106] C413 Satisfaction Rate. The satisfaction rate is the ratio of user waiting time to the maximum acceptable waiting time. This metric reflects situations where waiting times are too long.

[0107] C414 Maximum queue length. The maximum queue length is the maximum length of vehicles waiting in line at a service area at any given time. The longer the queue length, the greater the pressure on the charging service at the service area, and it is also an important indicator reflecting the service level.

[0108] C42 Charging Pile Standards. These standards, from a management perspective, aim to ensure that charging piles meet user charging needs while maintaining reasonable construction and maintenance costs.

[0109] C421 Charging Pile Construction and Maintenance Costs. Charging pile construction and maintenance costs are the sum of the construction and maintenance costs of each individual charging pile. This indicator reflects the cost level of charging piles within the service area.

[0110] C422 Charging Pile Utilization Rate. The charging pile utilization rate is the ratio of the charging pile's usage time to its available time. This indicator reflects the utilization status of charging piles in the service area.

[0111] Based on the technical concept described in the above method, this application also provides a system embodiment corresponding to the above method, the system comprising:

[0112] The multi-agent simulation module is used to build and run an interactive environment consisting of a highway agent, a vehicle agent, and a power dispatching system agent, wherein:

[0113] The highway agent is configured to dynamically update the road segment speed by calling the BPR function based on real-time traffic flow, and to provide route and speed information to the vehicle agent.

[0114] The vehicle agent is configured to update its location and battery status during driving, and to trigger a charging decision based on a mandatory charging threshold, an optional charging threshold, and probabilistic rules when entering a service area.

[0115] The power allocation system Agent is configured to dynamically allocate charging power requests at each level according to the two-level proportional allocation principle of "grid-service area-charging pile" under the constraint of the maximum available power Pmax of the power grid;

[0116] The optimization calculation module is used to run the particle swarm optimization algorithm with the goal of minimizing the weighted sum of the average user queuing time and the construction and operation cost of charging piles. It iteratively optimizes the number and type ratio of charging piles in each service area and outputs the optimal configuration scheme.

[0117] The data interface module is used to input traffic demand data, power grid parameters and charging pile parameters into the multi-agent simulation module, and to write the configuration scheme obtained by the optimization calculation module into an external database or scheduling system.

[0118] Example:

[0119] This embodiment covers the Zhejiang-Shanghai-Ningbo Expressway (Hangzhou-Ningbo section and Shanghai-Hangzhou section), with a total length of 248 kilometers. The route passes through four service areas: Yuyao Service Area, Shaoxing Service Area, Chang'an Service Area, and Jiaxing Service Area, in the order from Ningbo towards Hangzhou and Shanghai. A road diagram is shown below. Figure 1 As shown in Table 1, the toll station situation of the Shanghai-Hangzhou-Ningbo Expressway is shown in Table 2, and the service area situation is shown in Table 3.

[0120] Table 1. Toll Stations and Numbers of the Shanghai-Hangzhou-Ningbo Expressway

[0121]

[0122] Table 2 Service Areas and Numbers of the Shanghai-Hangzhou-Ningbo Expressway

[0123]

[0124] Based on the above actual situation, the steps of this embodiment are as follows:

[0125] 1. Constructing a road network agent

[0126] (1) Road network construction. The expressway framework is constructed with toll stations as nodes and roads as segments. The constructed expressway includes 20 toll stations, 4 service areas and 19 road segments. The node information is shown in Table 1 and the road segment information is shown in Table 3.

[0127] Table 3 Information on sections of the Shanghai-Hangzhou-Ningbo Expressway

[0128]

[0129] The completed road network topology map is as follows: Figure 2 As shown.

[0130] (2) Dynamic congestion feedback. When the vehicle agent queries the road segment speed, the road network agent calls the BPR function, taking the road segment design capacity (determined by the number of lanes and speed limit) and real-time traffic flow as input parameters to calculate the current actual speed.

[0131]

[0132] 2. Construct a power dispatching system agent

[0133] (1) The power grid and power distribution parameters are shown in Table 4.

[0134] Table 4 Power Grid and Power Distribution Setting Parameters

[0135]

[0136] (1) Physical location of service areas. The physical location of each service area is determined based on the road segment where the service area is located and its distance from the node, as shown in Table 5.

[0137] Table 5 Physical Location of Service Areas

[0138]

[0139] Road network topology map and the geographical location of service areas, such as Figure 3 As shown.

[0140] (2) Charging pile configuration in service areas. The number of charging piles in each service area and the configuration of each type of charging pile are shown in Table 2. The relevant parameters such as the cost and quantity limit of charging piles are shown in Table 6.

[0141] Table 6 Charging Pile Setting Parameters

[0142]

[0143] 3. Build a vehicle agent

[0144] (1) Vehicle data processing. The existing dataset consists of ETC gantry data from highway toll stations in Zhejiang Province. It contains data from various toll stations in Zhejiang Province, but only some of the toll stations shown in Table 1 are truly relevant to the selected case. It also includes some useless data such as toll serial numbers and payment methods, as well as some missing, incorrect, and invalid data. Therefore, the dataset needs to be cleaned and filtered first to obtain a set of data that can be directly input into the simulation model, as shown in Table 7.

[0145] Table 7 Processed Vehicle Data

[0146]

[0147] By statistically analyzing the origin and destination nodes of vehicles and their generation time, we can obtain the vehicle's OD value and the 24-hour traffic demand distribution, such as... Figure 4 As shown.

[0148] (2) Vehicle Initialization. Based on the vehicle's OD value and the 24-hour traffic demand distribution, the vehicle generator creates a corresponding number of vehicles to enter the road network and assigns them vehicle ID, origin and destination toll stations, vehicle type, and whether they are electrified. The initialization SoC of electric vehicles follows the rules. Distribution. Relevant parameters for electric vehicles are shown in Table 8.

[0149] Table 8 Relevant Parameters of Electric Vehicles

[0150]

[0151] (3) Vehicle charging decision. The vehicle charging decision parameters are shown in Table 9.

[0152] Table 9 Vehicle Charging Decision Parameter Settings

[0153]

[0154] 4. Simulation process

[0155] (1) Initialize the road network agent, service area agent, and preprocess ETC data. Simulate the operation results of the agent system for one day (24 hours).

[0156] (2) The vehicle generator generates vehicle entry points into the road network based on the OD value and 24-hour traffic demand distribution. The vehicle location and SoC are updated every 1 minute. When a vehicle passes through a service area, it enters the charging decision process. If the vehicle chooses to charge, it enters the service area charging service sequence. After charging is completed, the vehicle re-enters the road network and continues driving according to the planned route.

[0157] (3) Update service area status. Update the status of charging piles in the service area according to the vehicle charging selection, and record statistical elements such as charging demand and queuing time.

[0158] (4) End of run. The simulation ends when 24 hours have elapsed.

[0159] 5. Construction of Particle Swarm Optimization Algorithm

[0160] (1) Initialize the particle swarm. The particle count is set to 15, within the charging station range. The initial position of the particle is randomly generated, the initial velocity is 0, the optimal position of the individual is initialized to the current position, and the fitness is initialized to positive infinity.

[0161] (2) Fitness assessment. The average queuing time and charging pile cost of users are obtained through simulation. The fitness value is calculated to assess the fitness of the current particle. The fitness-related parameters are shown in Table 10.

[0162] Table 10 Fitness Parameter Settings

[0163]

[0164] (3) Update the optimal position. For each particle, if the fitness of the current particle is less than the fitness of the previous position, then update the individual's optimal position. If the fitness of the current particle is less than the fitness of the historical optimal position, then update the historical optimal position.

[0165] (4) Determine the number of iterations. The maximum number of iterations is 10. If the maximum number of iterations is reached, the iteration ends and the optimal configuration and performance indicators are output; if the maximum number of iterations is not reached, proceed to the next step.

[0166] (5) Update speed and position. Parameters , , To prevent particles from deviating too far from the maximum speed, the velocity is set to 5.0. The particle position is determined to meet the charging station range limit. Repeat steps (2), (3), and (4).

[0167] 6. Optimize performance index analysis

[0168] Optimize the configuration of front and rear charging stations, such as Figure 5 As shown, the optimization results indicate an increase in the number of charging stations in each service area and a higher proportion of overcharging. The convergence curve of the optimization algorithm is shown in Figure 1. Figure 6 As shown, with the increase of the number of iterations, the algorithm continuously finds solutions with smaller fitness.

[0169] (1) Average waiting time.

[0170] Results of average waiting times for each service area and the overall average waiting time are as follows: Figure 7 As shown, the average waiting time in each service area has decreased, and overall it has decreased by 45.5%.

[0171] (2) The proportion of those not queuing.

[0172] Results of the percentage of vehicles without queues at each service area and overall are as follows: Figure 8 As shown, the service area rate decreased slightly in Chang'an, increased slightly in Jiaxing, increased by 30.4% in Yuyao, and increased by 15.5% in Shaoxing, resulting in an overall increase of 6.1%.

[0173] (3) Satisfaction rate.

[0174] Results for each service area and the overall satisfaction rate are as follows: Figure 9 As shown, except for the Jiaxing service area where satisfaction decreased by 25.9%, all other service areas saw improvements, with an overall increase of 3.3%.

[0175] (4) Maximum queue length.

[0176] The results of the maximum queue length for each service area and the overall vehicle queue are as follows: Figure 10As shown, the maximum queue length in each service area has decreased significantly, from the current maximum queue length of 168 vehicles to the optimized maximum queue length of 96 vehicles.

[0177] (5) Cost of charging piles.

[0178] The results of optimizing charging pile costs before and after (operation time of 5 years) are as follows: Figure 11 As shown, due to the increase in the number of charging piles and the increase in the proportion of overcharging, both construction and operating costs have increased, with an overall increase of 15.2%, but the total cost is still within the maximum acceptable cost range.

[0179] (6) Utilization rate of charging piles.

[0180] The utilization rates of each service area and the overall charging piles are as follows: Figure 12 As shown, the utilization rate of charging stations in each service area has increased slightly, with an overall increase of 4%.

[0181] Taking all indicators into account, the overall system performance is compared to... Figure 13 As shown, overall vehicle queuing evaluation indicators have been improved, and while the cost of charging stations has increased, it remains within an acceptable range. Therefore, the optimization results meet the optimization objectives.

[0182] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0183] The specific embodiments described herein are merely illustrative examples illustrating the spirit of the invention. The above embodiments only express several implementation methods of the invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims. For those skilled in the art, multiple variations and improvements can be made without departing from the concept of the invention. Therefore, the scope of protection of this invention should be determined by the appended claims.

Claims

1. A method for configuring charging piles in highway service areas that integrates power allocation and multi-agent systems, characterized in that, Includes the following steps: S1. Construct a multi-agent simulation environment consisting of a highway agent, a vehicle agent, and a power dispatching system agent, where: The highway agent is used to provide a static road network topology including nodes, road segments and service areas, and dynamically updates the road segment speed using the BPR function based on real-time traffic flow. The vehicle agent is used to generate electric vehicle entities based on the OD matrix and update the location and SoC in real time during driving. When the vehicle enters the service area, it triggers a charging decision based on the mandatory charging threshold, the optional charging threshold and the probabilistic rules. The power allocation system Agent is used to dynamically allocate charging power requests in each service area according to the two-level proportional allocation principle of "grid-service area-charging pile" under the constraint of the maximum power available by the power grid Pmax. S2. To minimize the weighted sum of average user queuing time and charging pile construction and operation costs, an optimization objective function including a queuing time penalty term is established, and the particle swarm optimization algorithm is used to iteratively optimize the number and type ratio of charging piles in each service area. S3. Output a charging pile configuration scheme that satisfies the upper and lower limits of the number of charging piles in the service area and minimizes the objective function.

2. The method as described in claim 1, characterized in that, The BPR function is specifically: in, It is the average speed of the road segment. It is the free-flow speed of the road segment. It refers to the traffic flow of the road segment. It refers to the road section's traffic capacity. , It is the hysteresis coefficient.

3. The method as described in claim 1 or 2, characterized in that, The highway agent maintains the vehicle speed of the road segment in real time through a directed graph structure. The vehicle agent obtains the dynamic speed of the road segment in minutes and updates the location and battery level.

4. The method as described in claim 3, characterized in that, The vehicle agent also uses preset electric vehicle penetration rate parameters, combined with vehicle model differentiation settings, to determine the electrification of each generated vehicle. For vehicle agents determined to be electric vehicles, their electrical attributes will be automatically initialized. This includes the battery capacity matched to the vehicle model, energy consumption per 100 kilometers, and initial charge level.

5. The method as described in claim 1, characterized in that, The principle of the two-level proportional allocation is as follows: Power grid-service area level: If the total requested power is less than or equal to Pmax, then the power will be fully satisfied; otherwise, the basic power will be allocated first according to the proportion of requested power in each service area, and then the remaining power will be allocated in a secondary manner according to the number of vehicles currently queuing in the service area from high to low. Service Area - Charging Pile Level: Available power is allocated according to the power request ratio of each charging pile.

6. The method as described in claim 5, characterized in that, After the power allocation system Agent completes the power allocation at the grid-service area level, it broadcasts the actual available power of each service area to the charging pile Agent of the corresponding service area for allocation at the next level.

7. The method as described in claim 5 or 6, characterized in that, After each power reallocation, the power allocation system Agent writes the actual available power of each service area into a shared message queue, which the vehicle Agent can read in real time when selecting a charging station.

8. The method as described in claim 1, characterized in that, The specific optimization objective function is as follows: in This is the sum of the costs of charging stations. This represents the maximum acceptable cost for charging stations. Average user queuing time The maximum acceptable queuing time for user satisfaction. and As weight, This is a penalty item.

9. The method as described in claim 8, characterized in that, The penalty term P in the optimization objective function is triggered only when the average queuing time in any service area exceeds the maximum acceptable queuing time for users.

10. A charging pile configuration system for highway service areas, characterized in that, include: The multi-agent simulation module is used to build and run an interactive environment consisting of a highway agent, a vehicle agent, and a power dispatching system agent, wherein: The highway agent is configured to dynamically update the road segment speed by calling the BPR function based on real-time traffic flow, and to provide route and speed information to the vehicle agent. The vehicle agent is configured to update its location and battery status during driving, and to trigger a charging decision based on a mandatory charging threshold, an optional charging threshold, and probabilistic rules when entering a service area. The power allocation system Agent is configured to dynamically allocate charging power requests at each level according to the two-level proportional allocation principle of "grid-service area-charging pile" under the constraint of the maximum power available by the power grid Pmax; The optimization calculation module is used to run the particle swarm optimization algorithm with the goal of minimizing the weighted sum of the average user queuing time and the construction and operation cost of charging piles. It iteratively optimizes the number and type ratio of charging piles in each service area and outputs the optimal configuration scheme. The data interface module is used to input traffic demand data, power grid parameters and charging pile parameters into the multi-agent simulation module, and to write the configuration scheme obtained by the optimization calculation module into an external database or scheduling system.

Images