A refined matching control method and system for new energy charging stations and users based on multi-algorithm fusion

Abstract

The present invention discloses a method and system for fine-grained matching control of new energy charging stations and users based on multi-algorithm fusion, which relates to the technical field of coordinated control of energy storage systems, including: constructing a logical power topology diagram of charging pile resources, virtualizing and aggregating the power of discrete charging piles, and generating a logical power pool; generating a set of candidate matching paths in combination with the logical power pool based on the power requirements, time windows and battery parameters of user charging requests; optimizing and screening the candidate path set using a simulated annealing algorithm, and constructing it as an objective function to output an optimal scheduling path; dynamically allocating power resources in the logical power pool based on the optimal scheduling path, and controlling the matching connection between the charging pile and the user terminal; the present application virtualizes the power resources of the charging pile through the logical power pool, uniformly regulates the resources of various charging equipment from a global perspective, effectively balances the charging load, and realizes fine-grained resource scheduling.

Description

Technical Field

[0001] The present invention relates to the technical field of coordinated control of energy storage systems. More specifically, the present invention relates to a method and system for fine-grained matching control of new energy charging stations and users based on multi-algorithm fusion. Background Art

[0002] With the rapid development of new energy vehicles, the number of electric vehicles on the road is climbing year by year, driving a growing demand for new energy charging infrastructure. As the core hub supporting the operation of new energy vehicles, charging stations' service capabilities, dispatching efficiency, and intelligence are directly related to the user experience of electric vehicle users and the operational quality of urban energy systems.

[0003] Traditional new energy charging stations often use static or rule-driven scheduling methods, lacking unified modeling and dynamic management of charging pile power resources. This results in some charging piles being overutilized, creating congestion, while others are idle, resulting in low resource utilization efficiency and difficulty optimizing overall operational status. This present invention proposes a solution to this problem. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a method and system for fine-grained matching control of new energy charging stations and users based on multi-algorithm fusion. By constructing a logical power topology diagram and a logical power pool, the power resources of the charging piles are virtualized and abstracted. The system can uniformly control the resources of various charging equipment from a global perspective, effectively balance the charging load, and realize fine-grained resource scheduling.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A multi-algorithm fusion-based refined matching control method for new energy charging stations and users includes: constructing a logical power topology diagram of charging pile resources, virtualizing and aggregating the power of discrete charging piles, and generating a logical power pool; generating a set of candidate matching paths based on the power requirements, time windows, and battery parameters of user charging requests in combination with the logical power pool; using a simulated annealing algorithm to optimize and screen the set of candidate paths, and outputting the optimal scheduling path with the objective functions of minimizing user waiting time, maximizing total power utilization, and minimizing scheduling conflicts between paths; and dynamically allocating power resources in the logical power pool based on the optimal scheduling path to control the matching connection between charging piles and user terminals.

[0007] In a preferred embodiment, the logical power topology diagram of the charging pile resources is constructed, the power of discrete charging piles is virtualized and aggregated, and a logical power pool is generated, specifically as follows: the control plane node is obtained, and each charging pile is abstracted as a virtual power channel based on the real-time operation data of the data plane device; a central power controller is constructed on the control plane, which receives and processes VPC status information and generates a logical power topology diagram; and a dynamically updated logical power pool is constructed based on the logical power topology diagram and the set of schedulable channels.

[0008] In a preferred embodiment, the method of generating a candidate matching path set in combination with a logical power pool is as follows: obtaining user charging request data and defining a multi-objective scheduling function; preliminarily screening available VPCs based on the real-time status of the logical power pool; randomly arranging the screened VPC combinations to generate multiple preliminary charging paths; extracting chromosome feature data of each path and calculating the fitness in combination with the multi-objective scheduling function; performing selection, crossover, and mutation operations on the path population based on the fitness to generate a new path population, and re-evaluating based on the fitness to select the optimal path to enter the next generation, thereby obtaining a candidate matching path set.

[0009] In a preferred embodiment, the selection, crossover, and mutation operations are performed on the path population according to the fitness to generate a new path population, specifically as follows: based on the fitness evaluation results, a non-dominated sorting strategy is used to perform multi-objective sorting on the path individuals; the top K paths in the sorting are selected as parent paths, and the top n paths are screened as elite individuals; the VPC nodes with higher scheduling efficiency in the two parent paths are extracted as the crossover starting points; under the time window consistency constraint, new path segments are combined through position-preserving crossover to generate a crossover path, and the crossover path is integrity-checked; multi-granularity knowledge-guided mutation is applied to the crossover path after verification; the crossover path after verification is combined with the new path generated after mutation and the elite individuals to form a new generation population.

[0010] In a preferred embodiment, the simulated annealing algorithm is used to optimize and screen the candidate path set, with the objective functions of minimizing user waiting time, maximizing total power utilization, and minimizing scheduling conflicts between paths, and outputting the optimal scheduling path, specifically as follows: the path in the candidate matching path set is used as the initial solution of simulated annealing, and the objective function of path scheduling optimization is determined; the initial parameters of simulated annealing are set, the current path solution is initialized, and a neighborhood solution is constructed; according to the objective function of path scheduling optimization, the target value and target increment of the current path solution and the neighborhood solution are calculated; the Metropolis criterion is executed according to the target increment to perform solution acceptance judgment and update the current path solution; when the current path solution is better than the historical optimal solution, the global optimal path solution is updated; the current temperature parameter is updated according to the preset annealing function, and when the convergence condition is met, the iteration is terminated and the global optimal solution is recorded; the global optimal path solution is output as the final scheduling path between the new energy charging station and the user.

[0011] In a preferred embodiment, the current path solution is initialized and the neighborhood solution is constructed as follows: the current path solution is split into multiple path segments; based on the logical power topology graph, a subgraph covering the current path segment is selected from the logical power pool; attribute matching and substitutability determination are performed on each candidate VPC path in the subgraph to screen out candidate path segments that meet preset conditions; scheduling evaluation data is extracted for each candidate path segment, and the one with the best scheduling score is selected to replace the target path segment in the original path, thereby generating a neighborhood path solution.

[0012] In a preferred embodiment, the power resources in the logical power pool are dynamically allocated according to the optimal scheduling path, and the matching connection between the charging pile and the user terminal is controlled as follows: a dynamic scheduling token bound to the optimal scheduling path is generated, and the optimal scheduling path includes a power slice combination allocated in a time series; the power slice is marked as frozen in the logical power pool; a reservation confirmation request is pushed to the user terminal, and the scheduling token status is updated to locked according to user feedback; based on the locked scheduling token, the energy storage module and the power device are controlled to perform power buffering according to the scheduling path; after the vehicle enters the target parking space, the real-time scheduling engine is started, and the power slice usage and the physical status of the pile position are synchronously monitored; when the scheduling is detected to be completed or the user actively terminates it, the power slice resources are released and the scheduling token is set to an invalid state.

[0013] A dual-subject refined matching control system for new energy charging stations and users based on multi-algorithm fusion includes a power pool module, a candidate module, a screening module and a control module, and there are connections between the modules; the power pool module is used to construct a logical power topology diagram of charging pile resources, virtualize and aggregate the power of discrete charging piles, and generate a logical power pool; the candidate module is used to generate a set of candidate matching paths based on the power requirements, time window and battery parameters of the user's charging request in combination with the logical power pool; the screening module is used to optimize and screen the candidate path set using a simulated annealing algorithm, and output the optimal scheduling path with the objective functions of minimizing user waiting time, maximizing total power utilization and minimizing scheduling conflicts between paths; the control module is used to dynamically allocate power resources in the logical power pool according to the optimal scheduling path, and control the matching connection between charging piles and user terminals.

[0014] The technical effects and advantages of the present invention's method and system for fine-tuning the matching control of new energy charging stations and users based on multi-algorithm fusion are as follows:

[0015] 1. This invention virtualizes the power resources of traditional physical charging piles by constructing a logical power topology diagram, forming a dynamically manageable logical power pool. This design overcomes the limitations of direct scheduling of physical devices, achieving unified abstraction and centralized resource control, and providing a good foundation for subsequent path generation and scheduling optimization. This logical power pool can flexibly reflect the dispatchable status of charging equipment based on real-time data, greatly improving the system's visualization and responsiveness.

[0016] 2. This invention introduces a multi-objective evolutionary strategy, combined with multiple scheduling functions (such as energy utilization, waiting time, and device load balancing), to perform multiple rounds of evaluation and optimization of candidate paths. Through the selection, crossover, and mutation of individual paths, coupled with a non-dominated sorting and elite retention strategy, the system dynamically generates a set of candidate matching paths that meet these multiple objectives. This mechanism not only enhances path diversity but also provides a rich solution space for the final scheduling optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The figure is a flow chart of a dual-subject refined matching control method of a new energy charging station and a user based on multi-algorithm fusion according to the present invention.

[0018] Figure 2 This is a structural diagram of a dual-subject refined matching control system for new energy charging stations and users based on multi-algorithm fusion in the present invention. DETAILED DESCRIPTION

[0019] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0020] Example 1, Figure 1 The present invention provides a dual-subject refined matching control method of a new energy charging station and a user based on multi-algorithm fusion, including:

[0021] S1, builds a logical power topology map of charging pile resources, virtualizes and aggregates the power of discrete charging piles, and generates a logical power pool;

[0022] In this embodiment, a logical power topology diagram of charging pile resources is constructed, and the power of discrete charging piles is virtualized and aggregated to generate a logical power pool, as follows:

[0023] Physical charging piles, power distribution cabinets, power sensors, and other hardware devices within the charging station are uniformly regarded as data plane nodes, which only undertake energy flow and status collection functions. A central dispatch controller is introduced as a control plane node, responsible for issuing dispatch policies, managing power paths, and dynamically allocating power permissions, achieving separation of control and execution.

[0024] Based on the real-time operating data of data-plane devices, the control plane node abstracts each charging pile into a virtual power channel (VPC). The VPC contains the following attributes: maximum output power value, current available power, dynamic range for power adjustment, estimated release time window, current operating status (idle / occupied / faulty), and logical address and physical location information.

[0025] A central power controller is built on the control plane. This controller is deployed on an edge server or cloud platform. Based on the received VPC (virtual power channel) status information, it generates and maintains a logical power topology graph that describes the power flow relationships, channel availability, and load status of the entire site. The graph uses VPCs as edges and physical power units (such as charging piles, energy storage modules, and distribution units) as nodes. Node attributes, including adjustable bandwidth, status labels, and priority parameters, are embedded in the corresponding VPCs. The graph structure is generated based on the physical power connection relationships.

[0026] Based on the logical power topology and the set of schedulable channels, a logical power pool is constructed, in which the available power of each virtual power channel (VPC) is sliced ​​according to preset units, and the allocation period, scheduling priority and source channel identifier of each power slice are recorded.

[0027] S2, based on the power requirement, time window and battery parameters of the user's charging request, a set of candidate matching paths is generated in combination with the logical power pool;

[0028] In this embodiment, based on the power requirement, time window, and battery parameters of the user's charging request, a set of candidate matching paths is generated in combination with the logical power pool, as follows:

[0029] Obtaining user charging request data and defining a multi-objective scheduling function, wherein the multi-objective scheduling function includes minimizing charging time, minimizing power waste, and minimizing path conflict;

[0030] Preliminary screening of available charging pile resources, namely virtual power channels (VPCs), based on the current state of the logical power pool. The current state includes: being idle or about to be released; the dynamic power adjustment range covers the vehicle's allowed charging power; and the time window of the available power slice overlaps with the user's demand period.

[0031] The selected VPC combinations are randomly arranged to generate multiple preliminary charging paths. Each path represents a candidate charging solution and includes: the selected charging pile sequence (VPC identifier), the power value allocated to the vehicle (within the available VPC slice range), and the charging time period;

[0032] Encode the key parameters of each charging path (e.g., charging station ID, allocated power, start time) into a chromosome structure (e.g., binary string, real number vector);

[0033] Extract characteristic data from the chromosome and combine it with a multi-objective scheduling function to output a fitness evaluation result. The characteristic data includes charging time, power waste (efficiency of power resource allocation), and path conflict degree.

[0034] According to the fitness evaluation results, the path population is selected, crossed, and mutated to generate a new path population. The new path population is re-evaluated based on the fitness and the optimal path is selected to enter the next generation to obtain a set of candidate matching paths.

[0035] The fitness evaluation formula is as follows:

[0036]

[0037]

[0038]

[0039]

[0040] Where: is the fitness evaluation result, is the minimum satisfaction value of the three characteristic data, , is the interaction term of the three feature data, , is the charging time satisfaction function (the goal is to minimize the charging time), is the power waste satisfaction function (the goal is to minimize waste), is the path conflict satisfaction function (target minimum conflict), is the base of natural logarithms, is the interaction adjustment parameter, which controls the influence of the interaction term on the fitness. It is the charging time, is the power waste rate, is the path conflict degree, is the adjustment parameter of charging time, is the regulation parameter for power waste, It is the adjustment parameter of path conflict.

[0041] It should be noted that Determines the degree of influence of charging time on fitness, the larger This will make the effect of charging time on fitness more sensitive, that is, every increase in charging time will lead to a sharp drop in fitness. The function is designed to decay exponentially. The shorter the charging time, the higher the satisfaction. The exponential function can effectively simulate the "nonlinear" effect of time and avoid the errors of simple linear estimation.

[0042] Further, Determines the degree of penalty for power waste on fitness, the larger It will make the impact of power waste on fitness more significant. The function uses the square term to strengthen the penalty for large power waste, because power waste has a cumulative effect, and large power waste will have a greater impact on the overall efficiency of the system. The square term makes the fitness function more sensitive to systems with large power waste.

[0043] Furthermore, Determines the degree of influence of path conflict on fitness. This makes the impact of path conflict on fitness more significant. The function describes the impact of path conflict using a linear fraction. The greater the conflict, the faster the fitness decreases. Fitness is expressed in inverse form, so that when the path conflict is high, the fitness decreases more significantly.

[0044] Finally, the interaction term accounts for the interactions between different features. When multiple features perform poorly simultaneously, the fitness decreases nonlinearly, introducing a more refined penalty mechanism. By introducing the interaction term, the system can better handle the coupling effects between different factors. The final fitness is determined based on the worst-performing metric (charging time, power waste, and path conflict). The weighted interaction term enhances the nonlinear penalty, enabling the system to perform global optimization across multiple dimensions, preventing the performance of one metric from masking the shortcomings of other important metrics.

[0045] The path population is selected, crossed, and mutated according to the fitness evaluation results, as follows:

[0046] Based on the fitness evaluation results, a non-dominated sorting strategy is used to perform multi-objective sorting on the path individuals, and the crowding distance is combined to maintain diversity. According to the sorting results, the top K paths with the highest fitness are selected as the parent paths and passed to the next crossover operation. A portion of elite individuals are retained to directly enter the next generation to ensure that the good genes are not destroyed. The elite individuals are the top n paths selected from the top K paths with the highest fitness and retained.

[0047] Extract the VPC (virtual power channel) node with higher scheduling efficiency from the two parent paths as the intersection starting point;

[0048] Under the condition of time window consistency, new path segments are combined through position-preserving crossover or sequence-matching crossover to obtain crossover paths;

[0049] Perform integrity checks on cross paths. If any abnormalities such as logical power pool conflicts or time window overlaps occur, backup power slices are called or virtual power channel resources are re-matched for automatic repair.

[0050] Apply multi-granularity knowledge-guided mutation to the verified cross paths;

[0051] The verified cross paths and the new path individuals generated after mutation are combined with the retained elite individuals to form a new generation of population.

[0052] Applying multi-granularity knowledge-guided mutation to the verified cross paths specifically includes:

[0053] Node-level mutation: Replaces VPC nodes with excessive load or potential failures in the path with adjustable points with higher remaining power in the logical power pool.

[0054] Segment-level mutation: Replace high-conflict path segments to reduce system peak load;

[0055] Time window disturbance: fine-tuning the scheduling period to release peak power resources without affecting task completion;

[0056] Knowledge guidance mechanism: Based on historical operation data and genetic evolution effects, dynamically adjust the mutation probability and area to improve the effectiveness and convergence speed of mutation.

[0057] S3, uses the simulated annealing algorithm to optimize and screen the candidate path set, with the objective function of minimizing user waiting time, maximizing total power utilization and minimizing scheduling conflicts between paths, and outputs the optimal scheduling path;

[0058] In this embodiment, a simulated annealing algorithm is used to optimize and screen the candidate path set, with the objective functions of minimizing user waiting time, maximizing total power utilization, and minimizing scheduling conflicts between paths, and outputting the optimal scheduling path, as follows:

[0059] Using paths in the candidate matching path set as initial solutions for simulated annealing, and determining an objective function for path scheduling optimization, wherein the objective function for path scheduling optimization includes minimizing user waiting time, maximizing total power utilization, and minimizing scheduling conflicts between paths;

[0060] Set the initial parameters of simulated annealing and initialize the current path solution For the genetic algorithm output path, the simulated annealing initial parameters include setting the initial temperature , annealing factor , maximum number of iterations , neighborhood search times and the termination temperature ;

[0061] According to the current path after initialization , construct neighborhood solutions ;

[0062] According to the objective function of path scheduling optimization, output the current path solution and neighborhood solutions Target value and , and calculate the target increment ;

[0063] Execute the Metropolis criterion according to the target increment to determine the solution acceptance and update the current path solution, including:

[0064] like , then accept the neighborhood solution as the current path solution and update the current path solution;

[0065] like , then with probability Accept the neighborhood solution. If accepted, update the current path solution to the neighborhood solution. Otherwise, keep the original path solution unchanged. is the current temperature value;

[0066] When the current path solution is better than the historical optimal solution, update the global optimal path solution;

[0067] According to the preset annealing function Update current temperature ,in is the temperature of the system before this round of iteration, It is the temperature used for the next round of search after the current round ends, and records the number of acceptances and iterations in this round. When it meets or the maximum number of iterations has been exceeded If any of the conditions are met, convergence is determined, that is, the current path solution is better than the historical optimal solution, the iteration process ends, and it is recorded as the new global optimal solution. Otherwise, the process returns to the step of constructing a neighborhood solution and continues to generate a new neighborhood solution until convergence.

[0068] The final global optimal path solution is output as the final scheduling path between the new energy charging station and the user. The scheduling path includes the user's reserved charging pile sequence, the power slice information used, the scheduling time period, and the logical power pool resource identifier mapped thereto.

[0069] The target value The details are as follows:

[0070]

[0071] Where: is the scheduling conflict between paths, is the total power utilization, is the user waiting time, 、 and is the weight coefficient, is the target value of the current path solution.

[0072] The current path solution after initialization , construct neighborhood solutions , as follows:

[0073] Unpack the current path Split into multiple path segments (i.e., continuous virtual power channel pairs). Each segment contains one or more continuous VPC connection relationships, which are used to describe the scheduling relationship between a charging pile and its upstream and downstream power nodes.

[0074] Based on the logical power topology graph, a subgraph covering the current path segment is selected from the logical power pool. The subgraph includes all VPC nodes in the original path segment and the redundant VPC nodes directly connected to it.

[0075] Perform attribute matching and substitutability checks on each candidate VPC path in the subgraph, and select paths that meet the following conditions as candidate path segments:

[0076] Each VPC in the path exists in the logical power pool and has a status of "idle" or "schedulable";

[0077] Power slicing of all path segments supports continuous scheduling within the target reservation time window;

[0078] The path length or number of hops does not exceed the set threshold;

[0079] The nodes contained do not physically conflict with other high-priority tasks;

[0080] The scheduling evaluation data of each candidate path segment is extracted, and the one with the best scheduling score (such as high power utilization, low load, and low power loss) is selected to replace the target path segment in the original path. A new scheduling path solution is generated, which is recorded as the neighborhood path solution. The scheduling evaluation data includes power utilization, load, and power loss.

[0081] S4, according to the optimal scheduling path, dynamically allocates power resources in the logical power pool and controls the matching connection between the charging pile and the user terminal;

[0082] In this embodiment, according to the optimal scheduling path, the power resources in the logical power pool are dynamically allocated to control the matching connection between the charging pile and the user terminal, as follows:

[0083] Generate a scheduling token based on the optimal scheduling path, the scheduling token recording the scheduling task identifier, user identity information, power quota and scheduling time window;

[0084] Mark the power slices involved in the optimal scheduling path as frozen in the logical power pool to prevent the resources from being occupied by other scheduling tasks;

[0085] Push a reservation confirmation request to the user terminal, which includes the recommended charging pile number, scheduling time, power quota, parking space location information and route guidance information;

[0086] Receive the user's reservation confirmation feedback. If the user confirms, set the dispatch token status to "locked" and activate the geographic guidance equipment to implement vehicle path diversion and parking guidance. The guidance equipment includes ground projection, parking space lighting guidance and intelligent gate control;

[0087] Based on the locked scheduling token, the energy storage module and power equipment are controlled to prepare energy and buffer power according to the power slice information in the scheduling path, ensuring the stability of the power supply in the scheduling path.

[0088] After a vehicle enters the target parking space, the path scheduling task is initiated, and power usage and parking status are monitored in real time. If a path bottleneck or anomaly is detected, the dynamic path repair mechanism is triggered. Based on the logical power topology, the path segment is quickly replaced or the power slice configuration is adjusted to ensure the continuous execution of the scheduling task.

[0089] When the scheduling task is completed or the user ends the task ahead of schedule, the occupied power resources are released, the scheduling token status is set to "invalid", and the scheduling log is recorded for subsequent credit evaluation and system optimization.

[0090] Example 2, Figure 2 The present invention provides a dual-subject refined matching control system for new energy charging stations and users based on multi-algorithm fusion, including a power pool module, a candidate module, a screening module, and a control module, with connections between the modules.

[0091] The power pool module is used to build a logical power topology diagram of charging pile resources, virtualize and aggregate the power of discrete charging piles, and generate a logical power pool;

[0092] The candidate module is used to generate a set of candidate matching paths based on the power requirements, time window and battery parameters of the user's charging request and the logical power pool;

[0093] The screening module is used to optimize and screen the candidate path set using the simulated annealing algorithm, with the objective functions of minimizing user waiting time, maximizing total power utilization, and minimizing scheduling conflicts between paths, and outputs the optimal scheduling path;

[0094] The control module is used to dynamically allocate power resources in the logical power pool according to the optimal scheduling path and control the matching connection between the charging pile and the user terminal.

[0095] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0096] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0097] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0098] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0099] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0100] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A dual-subject refined matching control method for new energy charging stations and users based on multi-algorithm fusion, characterized by: include: Build a logical power topology diagram of charging pile resources, virtualize and aggregate the power of discrete charging piles, and generate a logical power pool; The logical power topology diagram of the charging pile resources is constructed, and the power of discrete charging piles is virtualized and aggregated to generate a logical power pool. Specifically: Obtain control plane nodes and abstract each charging pile into a virtual power channel based on the real-time operating data of data plane devices; Build a central power controller on the control plane to receive and process VPC status information and generate a logical power topology map; Build a dynamically updated logic power pool based on the logic power topology graph and the set of schedulable channels; Based on the power requirements, time window, and battery parameters of the user's charging request, a set of candidate matching paths is generated in combination with the logical power pool; The generation of a candidate matching path set by combining the logic power pool is specifically as follows: Obtain user charging request data and define a multi-objective scheduling function; Preliminary screening of available VPCs based on the real-time status of the logical power pool; Randomly arrange the selected VPC combinations to generate multiple preliminary charging paths; Extract the chromosome feature data of each path and calculate the fitness by combining it with the multi-objective scheduling function; According to the fitness, the path population is selected, crossed, and mutated to generate a new path population. The fitness is re-evaluated and the optimal path is selected to enter the next generation to obtain a set of candidate matching paths. The simulated annealing algorithm is used to optimize and screen the candidate path set, with the objective function of minimizing user waiting time, maximizing total power utilization and minimizing scheduling conflicts between paths, and outputting the optimal scheduling path. The simulated annealing algorithm is used to optimize and screen the candidate path set, with the objective function of minimizing user waiting time, maximizing total power utilization, and minimizing scheduling conflicts between paths, and outputting the optimal scheduling path, specifically: The paths in the candidate matching path set are used as the initial solutions of simulated annealing, and the objective function of path scheduling optimization is determined; Set the initial parameters of simulated annealing, initialize the current path solution, and construct the neighborhood solution; According to the objective function of path scheduling optimization, the target value and target increment of the current path solution and the neighborhood solution are calculated; According to the target increment, the Metropolis criterion is executed to judge the solution acceptance and update the current path solution; When the current path solution is better than the historical optimal solution, update the global optimal path solution; Update the current temperature parameter according to the preset annealing function, terminate the iteration when the convergence condition is met and record the global optimal solution; Output the global optimal path solution as the final scheduling path between the new energy charging station and the user; Dynamically allocate power resources in the logical power pool based on the optimal scheduling path to control the matching connection between charging piles and user terminals; The method of dynamically allocating power resources in the logical power pool according to the optimal scheduling path and controlling the matching connection between the charging pile and the user terminal is as follows: generating a dynamic scheduling token bound to an optimal scheduling path, wherein the optimal scheduling path includes a combination of power slices allocated in a time series; Marking the power slice as frozen in the logical power pool; Push the reservation confirmation request to the user terminal and update the scheduling token status to locked based on the user feedback; Based on the locked scheduling token, the energy storage module and power equipment are controlled to perform power buffering according to the scheduling path; After the vehicle enters the target parking space, the real-time scheduling engine is activated to simultaneously monitor the power slice usage and the physical status of the parking space; When the scheduling is completed or the user actively terminates the scheduling, the power slice resources are released and the scheduling token is set to an invalid state.

2. The multi-algorithm fusion-based dual-agent refined matching control method for new energy charging stations and users according to claim 1 is characterized in that: The selection, crossover, and mutation operations are performed on the path population according to the fitness to generate a new path population, as follows: Based on the fitness evaluation results, a non-dominated sorting strategy is used to sort the path individuals by multiple objectives. Select the top K paths as parent paths, and select the top n paths as elite individuals; The VPC node with higher scheduling efficiency in the two parent paths is selected as the intersection starting point. Under the time window consistency constraint, new path segments are combined by position-preserving cross-combination to generate a cross-path, and the integrity of the cross-path is checked. Apply multi-granularity knowledge-guided mutation to the verified cross paths; The verified cross paths are combined with the new paths generated after mutation and elite individuals to form a new generation of population.

3. The multi-algorithm fusion-based refined matching control method for new energy charging stations and users according to claim 1 is characterized in that: Initializing the current path solution and constructing the neighborhood solution are as follows: Split the current path into multiple path segments; Based on the logical power topology graph, a subgraph covering the current path segment is selected from the logical power pool; Perform attribute matching and substitutability determination on each candidate VPC path in the subgraph to select candidate path segments that meet preset conditions. The scheduling evaluation data of each candidate path segment is extracted, and the one with the best scheduling score is selected to replace the target path segment in the original path to generate a neighborhood path solution.

4. A system using the multi-algorithm fusion-based dual-agent refined matching control method for new energy charging stations and users as described in any one of claims 1 to 3, characterized in that: It includes power pool module, candidate module, screening module and control module, and there are connections between modules; The power pool module is used to build a logical power topology diagram of charging pile resources, virtualize and aggregate the power of discrete charging piles, and generate a logical power pool; The candidate module is used to generate a set of candidate matching paths based on the power requirements, time window and battery parameters of the user's charging request and the logical power pool; The screening module is used to optimize and screen the candidate path set using the simulated annealing algorithm, with the objective functions of minimizing user waiting time, maximizing total power utilization, and minimizing scheduling conflicts between paths, and outputs the optimal scheduling path; The control module is used to dynamically allocate power resources in the logical power pool according to the optimal scheduling path and control the matching connection between the charging pile and the user terminal.

Images