Wireless charging power adaptive matching method based on multi-objective optimization

Abstract

This invention discloses an adaptive power matching method for wireless charging based on multi-objective optimization, specifically relating to the field of wireless charging technology. By introducing two complexity evaluation features—Pareto conflict density and local dominance index—a sorting complexity index model is constructed. This model dynamically determines the sorting load and triggers a complexity reduction mechanism under high load conditions, effectively reducing non-dominated sorting overhead. While ensuring solution quality and diversity, the real-time performance and deployment feasibility of the algorithm in embedded wireless charging systems are improved, ultimately achieving efficient adaptive power matching control for different devices and environments.

Description

Technical Field

[0001] This invention relates to the field of wireless charging technology, and more specifically to a wireless charging power adaptive matching method based on multi-objective optimization. Background Technology

[0002] Wireless charging power adaptive matching refers to the ability of a charging device (such as a charger) to automatically adjust its output power based on the power requirements and operating status of the device being charged (such as a mobile phone or headphones) during wireless charging, in order to achieve optimal charging efficiency and safety. This technology can avoid energy waste, device overheating, or slow charging caused by power mismatch, thereby improving the overall user experience.

[0003] The existing technology has the following shortcomings: In existing technologies, under high-dimensional optimization scenarios (such as those involving four or more objectives like efficiency, temperature, lifetime, and harmonics), the non-dominated sorting step of the NSGA-II algorithm faces a sharp increase in computational complexity, with a time complexity of O(n log n). Where M is the target number and N is the population size. This computational explosion is particularly severe in embedded or real-time wireless charging systems, causing the response time of optimization algorithms to fail to meet millisecond-level control requirements, thus limiting their deployment and application in practical systems. Summary of the Invention

[0004] The purpose of this invention is to provide a wireless charging power adaptive matching method based on multi-objective optimization to address the shortcomings of the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a wireless charging power adaptive matching method based on multi-objective optimization, comprising: Collect various key performance parameters of the wireless charging system and construct a multi-objective optimization model; An initial population is constructed based on a non-dominated sorting genetic algorithm, and the fitness value of each individual in the multi-objective space is calculated. Extract feature parameters for complexity assessment from the current population, including Pareto conflict density and local dominance index; The Pareto conflict density and local dominance index are comprehensively calculated based on the preset complexity assessment model to obtain the sorting complexity index value, which is then compared with the preset threshold to determine whether there is a risk of a surge in sorting complexity. If the sorting complexity index value exceeds the preset threshold, a complexity reduction mechanism is triggered, including performing local non-dominated sorting on regions with high local dominance index and adopting a fast pre-screening strategy on regions with low local dominance index. Based on the improved sorting results, tournament selection, crossover, and mutation operations are performed to generate a new generation of individuals, and the optimization process is iteratively executed. The system outputs a dynamically adjusted and optimized Pareto optimal solution set and configures the selected power matching parameters into the wireless charging system to achieve an adaptive charging strategy for different devices and environments.

[0006] Preferably, the acquisition of multiple key performance parameters of the wireless charging system includes: real-time acquisition of the transmitter input power and receiver output power through a dual-channel voltage / current sampler to calculate charging efficiency parameters; monitoring of energy loss of key components in the system through thermistors and power sensors; acquisition of temperature rise data by deploying temperature sensors at heat-generating locations; and frequency domain analysis of the current signal using a high-speed ADC combined with an FFT module to obtain harmonic interference indicators.

[0007] Preferably, the optimization objectives in the multi-objective optimization model include: maximizing charging efficiency, minimizing energy loss, minimizing temperature rise, and minimizing harmonic interference.

[0008] Preferred, individual The fitness is a set of multiple objective function values: ;in: Let be the i-th individual, corresponding to a wireless charging parameter combination, and M be the number of objective functions in the multi-objective optimization. For the j-th objective function in individual The value to be taken below; for two individuals and If it satisfies all objective functions And there exists at least one target. Then it is believed that non-dominance .

[0009] The preferred Pareto conflict density extraction method is as follows: The current population is set as... Each individual has a corresponding function value for M objectives. For any pair of targets Calculate the conflict density: In the formula, Let be the index of the degree of conflict between the p-th objective function and the q-th objective function in the current population; R is the number of individuals in the current population. For individuals The function value on the objective function p, For individuals The function value on the objective function q; For individuals Relative to the changing trends of its neighboring individuals on target p, sign is the sign function. Averaging over all target pairs yields the Pareto conflict density (PCD), expressed as: .

[0010] The preferred method for extracting the local dominance index is as follows: for each individual Calculate the k nearest neighbors in the target space; count the number of neighbors in the neighborhood. satisfy non-dominance Define the local dominance index of an individual. The calculation expression is: The average of the entire population yields the Local Dominance Index (LDI), expressed as follows: In the formula, S represents the total number of individuals.

[0011] Preferably, the Pareto conflict density and local dominance index are converted into a comprehensive feature vector. The comprehensive feature vector is used as the input of the machine learning model. The machine learning model uses the prediction of the ranking complexity index label for each set of comprehensive feature vectors as the prediction objective and the minimization of the sum of prediction errors for all ranking complexity index labels as the training objective. The machine learning model is trained until the sum of prediction errors converges and the model training stops. The ranking complexity index value is determined based on the model output. The machine learning model is a multinomial regression model.

[0012] Preferably, the obtained sorting complexity index value is compared with a preset threshold. If the sorting complexity index value is greater than or equal to the preset threshold, it is considered that the current population has a risk of a surge in sorting complexity, and the corresponding complexity reduction processing mechanism is triggered. If the sorting complexity index value is less than the preset threshold, it is considered that the current sorting load is within a controllable range, and the standard global non-dominated sorting process continues to be executed.

[0013] Preferably, if it is determined that there is a risk of a surge in sorting complexity, the complexity reduction mechanism includes: Based on the local dominance index, the current population is divided into regions with high local dominance index and regions with low local dominance index; Perform precise non-dominated sorting on regions with high local dominance indices; For regions with low local dominance index, a fast heuristic pre-sorting strategy is adopted, including single-objective mapping sorting or weighted scoring sorting. The sorting results of the two regions are merged for subsequent genetic operations.

[0014] Preferably, the genetic operation includes: performing tournament selection based on the sorting results to select individuals with high fitness as parents; performing crossover on the selected parents according to a preset probability to generate offspring individuals; perturbing the parameters of the offspring individuals according to a preset probability to complete the mutation operation; and merging the offspring with the selected parents to form the next generation population.

[0015] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention introduces Pareto conflict density and local dominance index as complexity evaluation features to effectively quantify the computational pressure of non-dominated sorting. Combined with a machine learning model, it predicts sorting complexity, enabling dynamic judgment and management of sorting computation load. When excessive computational load is detected, this invention further employs a partitioning strategy, using locally precise sorting and rapid pre-screening respectively, to reduce overall computational complexity. This solves the problem of untimely computational response in traditional NSGA-II algorithms in embedded or real-time systems.

[0016] 2. This invention also combines the optimized population results with genetic operations and iterative updates to ultimately output a Pareto optimal solution set suitable for practical application scenarios. The selected power matching parameters are then configured into the wireless charging system, achieving efficient adaptive charging control for various types of devices and complex environments. The overall method not only improves optimization efficiency and system real-time performance but also possesses excellent engineering deployment capabilities, making it suitable for practical scenarios requiring high precision, low latency, and multi-objective coordination in wireless charging systems. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0018] Figure 1 This is a mind map of the method of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments 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.

[0020] For examples, please refer to Figure 1 As shown in this embodiment, the wireless charging power adaptive matching method based on multi-objective optimization includes: Collect various key performance parameters of the wireless charging system and construct a multi-objective optimization model; An initial population is constructed based on a non-dominated sorting genetic algorithm, and the fitness value of each individual in the multi-objective space is calculated. Extract feature parameters for complexity assessment from the current population, including Pareto conflict density and local dominance index; The Pareto conflict density and local dominance index are comprehensively calculated based on the preset complexity assessment model to obtain the sorting complexity index value, which is then compared with the preset threshold to determine whether there is a risk of a surge in sorting complexity. If the sorting complexity index value exceeds the preset threshold, a complexity reduction mechanism is triggered, including performing local non-dominated sorting on regions with high local dominance index and adopting a fast pre-screening strategy on regions with low local dominance index. Based on the improved sorting results, tournament selection, crossover, and mutation operations are performed to generate a new generation of individuals, and the optimization process is iteratively executed. The system outputs a dynamically adjusted and optimized Pareto optimal solution set and configures the selected power matching parameters into the wireless charging system to achieve an adaptive charging strategy for different devices and environments.

[0021] In this embodiment, the wireless charging system includes a transmitter and a receiver, and the system structure includes a power transmission module, a magnetically coupled resonant circuit, a rectifier and filter module, and a system controller.

[0022] The charging efficiency parameter η is measured using a dual-channel voltage / current sampler to measure the input power at the transmitter. With the output power of the receiver Real-time calculation of charging efficiency: The sampling period is set to every 10 ms, and the average efficiency and instantaneous fluctuation are calculated using a continuous sliding window.

[0023] Energy loss parameters The power loss of each component (coil, inductor, capacitor, MOSFET) is monitored in real time using precision thermistors and resistive power sensors. ;in The equivalent series resistance of the component. For the voltage drop of the switching device, The current flowing through it.

[0024] The temperature rise control parameter ΔT is set by placing NTC thermistors or DS18B20 temperature sensors at key locations on both the transmitter and receiver sides (transformer, MOSFET, coil center). A reference temperature is then set. Real-time collection of current operating temperature The temperature rise was then calculated. The temperature rise trend is assessed by combining the time function to reflect the effect of continuous heat accumulation.

[0025] The harmonic interference parameter THD was calculated by sampling the AC current signal on the transmitting side using a high-speed ADC in conjunction with a Fast Fourier Transform (FFT) module. The total harmonic distortion (THD) rate was then calculated. ;in For the nth harmonic current, For the fundamental frequency component, the range of 2kHz to 200kHz is usually selected, and N is the upper limit of the harmonic order (e.g., 20).

[0026] Based on the above key performance indicators, the following set of multi-objective functions is established: Maximize charging efficiency: ; Minimize energy loss: ; Minimize temperature rise level: ; Minimize harmonic distortion: ; Here, variable x represents the optimized system parameter vector, including but not limited to power output level, switching frequency, matching capacitor value, coil spacing, modulation duty cycle, etc.

[0027] The system controller uses an STM32H7 or higher performance embedded chip; all acquired signals are input through an isolated ADC (such as ADS131M04) and transferred to the embedded host via DMA; FFT acceleration calculation is performed using FPGA coprocessor; all parameters are normalized to 32-bit floating-point format for easy interface with NSGA-II algorithm.

[0028] This implementation method can systematically collect multi-dimensional key parameters reflecting wireless charging performance, construct an accurate, adjustable, and real-time iterative multi-objective optimization model, and provide complete input support for subsequent adaptive power matching based on NSGA-II.

[0029] In NSGA-II, each individual The fitness is not a single numerical value, but a set of multiple objective function values: ;in: Let be the i-th individual, corresponding to a wireless charging parameter combination, and M be the number of objective functions in the multi-objective optimization. For the j-th objective function in individual The fitness evaluation takes values ​​such as efficiency, loss, and temperature rise. Instead of directly synthesizing a single value, the fitness evaluation is achieved through two key processes: non-dominated sorting and crowding distance calculation.

[0030] For two individuals and If the following conditions are met: On all objective functions And there exists at least one target. Then it is believed that non-dominance .

[0031] The expression for calculating congestion distance is: Where M is the target total number; , For individuals The neighboring sorted individual values ​​on the j-th target, The maximum / minimum value of the j-th objective (used for normalization); the larger the crowding distance, the more "sparse" the solution is in the solution space, which helps to preserve diversity.

[0032] Pareto conflict density is used to quantify the degree of conflict between multiple objectives in the current population, that is, whether the improvement of one objective function is often accompanied by the deterioration of other objectives, reflecting the non-cooperation of the optimization space.

[0033] The extraction method is as follows: The current population is set as... Each individual has a corresponding function value for M objectives. ; For any pair of targets Calculate the conflict density: In the formula, Let R be the conflict density between the p-th objective function and the q-th objective function in the current population; R is the number of individuals in the current population. For individuals The function value on the objective function p, For individuals The function value on the objective function q; For individuals The trend of change of its neighboring individuals on target p can be approximated as follows: That is, the first-order difference; sign is the sign function, defined as: The Pareto conflict density (PCD) is obtained by averaging over all target pairs, and its expression is: In the formula, PCD∈[0,1], approaching 1 indicates high conflict and high pressure for multi-objective optimization; approaching 0 indicates cooperation between objectives, which is suitable for conventional non-dominated sorting.

[0034] The local dominance index is used to measure the degree to which a solution is non-dominated by other individuals in its neighborhood, and is used to identify the spatial distribution of non-dominated ordination pressure in the population.

[0035] The method for extracting the local dominant index is as follows: For each individual Calculate its k nearest neighbors in the target space (k=10 is usually chosen). How many neighbors are there within the neighborhood? satisfy non-dominance ; Define the local dominance index of an individual The calculation expression is: The average of the entire population yields the Local Dominance Index (LDI), expressed as follows: In the formula, S represents the total number of individuals.

[0036] The obtained overall Local Dominance Index (LDI) is compared with a preset Local Dominance Index reference threshold. If the LDI is greater than or equal to the reference threshold, it indicates that the non-dominated sorting operation is mainly concentrated in the local subspace of the population, resulting in excessively high local computational load. In this case, it is suitable to trigger a sorting alternative strategy based on local regions to reduce the overall computational complexity. If the LDI is less than the reference threshold, it indicates that the non-dominated relations are relatively evenly distributed in the population, and the sorting operation load is dispersed. In this case, the standard global non-dominated sorting operation can continue to be performed to maintain the comprehensive exploration and diversity of solutions.

[0037] Pareto conflict density and local dominance index are converted into a comprehensive feature vector. This comprehensive feature vector is used as input to a machine learning model. The machine learning model uses the prediction of ranking complexity index labels for each set of comprehensive feature vectors as its prediction objective and minimizes the sum of prediction errors for all ranking complexity index labels as its training objective. The machine learning model is trained until the sum of prediction errors converges, at which point the model training stops. The ranking complexity index value is determined based on the model output. The machine learning model is a multinomial regression model.

[0038] The obtained sorting complexity index value is compared with a preset threshold. If the sorting complexity index value is greater than or equal to the preset threshold, it is considered that the current population is at risk of a surge in sorting complexity, and the corresponding complexity reduction mechanism (such as local sorting, target compression, etc.) is triggered. If the sorting complexity index value is less than the preset threshold, it is considered that the current sorting load is within a controllable range, and the standard global non-dominated sorting process continues to be executed.

[0039] When the sorting complexity index value is greater than or equal to a preset threshold, a complexity reduction mechanism is triggered, including performing local non-dominated sorting on regions with high local dominance index and adopting a fast pre-screening strategy on regions with low local dominance index. Specifically: If the Local Dominance Index (LDI) is greater than or equal to the Local Dominance Index Reference Threshold, it is marked as a high Local Dominance Index region; if the Local Dominance Index (LDI) is less than the Reference Threshold, it is marked as a low Local Dominance Index region.

[0040] For individuals within high LDI regions, precise NSGA-II nondominated ordination is performed separately to ensure that optimal solutions in locally dense regions can be accurately identified, preventing degradation or false screening.

[0041] For individuals in low LDI regions, skip the complete non-dominated sorting process; instead, use a fast heuristic pre-sorting strategy (such as single-objective projection or weighted scoring) to select the top-ranked individuals; reduce redundant sorting calculations and improve overall operating efficiency.

[0042] Merge the sorting results of high LDI regions with the pre-sorting results of low LDI regions; Perform subsequent genetic operations (crossover, mutation, etc.) such as tournament selection; generate the next generation population, and continue to enter the NSGA-II iteration cycle.

[0043] From the improved ranking results, multiple candidate individuals are randomly selected based on their non-dominance level and crowding information. Within each group of candidate individuals, individuals with higher fitness are prioritized for retention based on ranking priority and distribution diversity. The selected individuals are used to form the parent population, providing a genetic basis for subsequent operations.

[0044] Select several pairs of individuals from the parent population and perform crossover operations according to a set probability. The parameter information of two individuals is combined to generate a new individual, thereby forming a new parameter region for exploration in the solution space; This ensures that crossover operations do not disrupt population diversity while enhancing the ability to search for the global optimum.

[0045] In the newly generated individuals, a small-scale perturbation is performed on certain parameters with a certain probability; Mutation operations help to escape local optima and improve the overall search capability of the population; Ensure that the variation range is within the system's safety range, and avoid unreasonable parameters causing anomalies in subsequent simulations or system operation.

[0046] The offspring individuals generated through crossover and mutation are merged with some of the elite individuals from the parent generation to form the next generation population. Repeatedly execute the optimization process, including multi-objective function evaluation, sorting complexity judgment, complexity reduction processing, selection and updating; The process is executed iteratively until a termination condition is met, such as the number of iterations reaching a set upper limit or the population convergence meeting a set standard.

[0047] In the final generation of the population, solutions located at the non-dominated front are selected to form the Pareto optimal solution set; This solution set reflects the optimal trade-offs among multiple objectives, covering different optimization directions such as high efficiency, low loss, low temperature rise, and low harmonics.

[0048] Based on actual application requirements (such as equipment type, ambient temperature, power supply conditions, etc.), select a set of power matching parameters that best fit the current strategy from the Pareto optimal solution set; The selected parameters are sent to the wireless charging system control module to dynamically configure the output power, modulation frequency, matching network parameters, etc. It enables adaptive and efficient wireless charging control for different devices and operating environments.

[0049] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0050] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0051] It should be understood that the term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Additionally, the character " / " in this document generally indicates an "or" relationship between the preceding and following related objects, but it may also indicate an "and / or" relationship; please refer to the context for specific understanding. Those skilled in the art will recognize that the units and algorithm steps of the various examples 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 implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art 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.

[0052] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A wireless charging power adaptive matching method based on multi-objective optimization, characterized in that: include: Collect various key performance parameters of the wireless charging system and construct a multi-objective optimization model; An initial population is constructed based on a non-dominated sorting genetic algorithm, and the fitness value of each individual in the multi-objective space is calculated. Extract feature parameters for complexity assessment from the current population, including Pareto conflict density and local dominance index; The Pareto conflict density and local dominance index are comprehensively calculated based on the preset complexity assessment model to obtain the sorting complexity index value, which is then compared with the preset threshold to determine whether there is a risk of a surge in sorting complexity. If the sorting complexity index value exceeds the preset threshold, a complexity reduction mechanism is triggered, including performing local non-dominated sorting on regions with high local dominance index and adopting a fast pre-screening strategy on regions with low local dominance index. Based on the improved sorting results, tournament selection, crossover, and mutation operations are performed to generate a new generation of individuals, and the optimization process is iteratively executed. The system outputs a dynamically adjusted and optimized Pareto optimal solution set and configures the selected power matching parameters into the wireless charging system to achieve an adaptive charging strategy for different devices and environments.

2. The wireless charging power adaptive matching method based on multi-objective optimization according to claim 1, characterized in that: The acquisition of various key performance parameters of the wireless charging system includes: real-time acquisition of the transmitter input power and receiver output power through a dual-channel voltage / current sampler to calculate charging efficiency parameters; monitoring of energy loss of key components in the system through thermistors and power sensors; acquisition of temperature rise data by deploying temperature sensors at heat-generating locations; and frequency domain analysis of the current signal using a high-speed ADC combined with an FFT module to obtain harmonic interference indicators.

3. The wireless charging power adaptive matching method based on multi-objective optimization according to claim 2, characterized in that: The optimization objectives in the multi-objective optimization model include: maximizing charging efficiency, minimizing energy loss, minimizing temperature rise, and minimizing harmonic interference.

4. The wireless charging power adaptive matching method based on multi-objective optimization according to claim 1, characterized in that: individual The fitness is a set of multiple objective function values: ;in: Let be the i-th individual, corresponding to a wireless charging parameter combination, and M be the number of objective functions in the multi-objective optimization. For the j-th objective function in individual The value to be taken below; for two individuals and If it satisfies all objective functions And there exists at least one target. Then it is believed that non-dominance .

5. The wireless charging power adaptive matching method based on multi-objective optimization according to claim 4, characterized in that: The Pareto conflict density extraction method is as follows: The current population is set as... Each individual has a corresponding function value for M objectives. For any pair of targets Calculate the conflict density: In the formula, Let be the index of the degree of conflict between the p-th objective function and the q-th objective function in the current population; R is the number of individuals in the current population. For individuals The function value on the objective function p, For individuals The function value on the objective function q; For individuals Relative to the changing trends of its neighboring individuals on target p, sign is the sign function. Averaging over all target pairs yields the Pareto conflict density (PCD), expressed as: .

6. The wireless charging power adaptive matching method based on multi-objective optimization according to claim 5, characterized in that: The method for extracting the local dominance index is as follows: for each individual Calculate the k nearest neighbors in the target space; count the number of neighbors in the neighborhood. satisfy non-dominance Define the local dominance index of an individual. The calculation expression is: The average of the entire population yields the Local Dominance Index (LDI), expressed as follows: In the formula, S represents the total number of individuals.

7. The wireless charging power adaptive matching method based on multi-objective optimization according to claim 6, characterized in that: Pareto conflict density and local dominance index are converted into a comprehensive feature vector. This comprehensive feature vector is used as input to a machine learning model. The machine learning model uses the prediction of ranking complexity index labels for each set of comprehensive feature vectors as its prediction objective and minimizes the sum of prediction errors for all ranking complexity index labels as its training objective. The machine learning model is trained until the sum of prediction errors converges, at which point the model training stops. The ranking complexity index value is determined based on the model output. The machine learning model is a multinomial regression model.

8. The wireless charging power adaptive matching method based on multi-objective optimization according to claim 7, characterized in that: The obtained sorting complexity index value is compared with a preset threshold. If the sorting complexity index value is greater than or equal to the preset threshold, it is considered that the current population is at risk of a surge in sorting complexity, and the corresponding complexity reduction mechanism is triggered. If the sorting complexity index value is less than the preset threshold, it is considered that the current sorting load is within a controllable range, and the standard global non-dominated sorting process continues to be executed.

9. The wireless charging power adaptive matching method based on multi-objective optimization according to claim 8, characterized in that: If it is determined that there is a risk of a surge in sorting complexity, the complexity reduction mechanism includes: Based on the local dominance index, the current population is divided into regions with high local dominance index and regions with low local dominance index; Perform precise non-dominated sorting on regions with high local dominance indices; For regions with low local dominance index, a fast heuristic pre-sorting strategy is adopted, including single-objective mapping sorting or weighted scoring sorting. The sorting results of the two regions are merged for subsequent genetic operations.

10. The wireless charging power adaptive matching method based on multi-objective optimization according to claim 9, characterized in that: The genetic operations include: performing tournament selection based on the sorting results to select individuals with high fitness as parents; performing crossover on the selected parents according to a preset probability to generate offspring individuals; perturbing the parameters of the offspring individuals according to a preset probability to complete the mutation operation; and merging the offspring with the selected parents to form the next generation population.

Images