Improved grey wolf optimizer-based electric vehicle charging station locating and sizing method
By introducing nonlinear convergence factor and weight update formula in the Gray Wolf optimization algorithm, the problem of local optimal and slow convergence speed in the site selection and capacity setting of electric vehicle charging stations is solved, and a more efficient charging station location is achieved and capacity configuration improves the uniformity and convenience of the charging network.
Patent Information
- Application Number
- CN202411005617.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-15
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The traditional gray wolf optimization algorithm is prone to falling into the local optimal solution and slow convergence speed in the site selection and capacity setting of electric vehicle charging stations, and it is difficult to find the optimal position and optimal capacity in high-dimensional space, resulting in unevenness and non-uniformity of the charging network. Convenience.
A nonlinear convergence factor and weighted position update formula are introduced, combined with multiple search strategies, candidate solutions are generated and optimal solutions are selected through fitness evaluation, enhancing the algorithm's global search ability and convergence speed.
It significantly improves the search efficiency and convergence speed of the algorithm, and can find a better charging station site selection and capacity setting solution within a reasonable time, avoid local optimal traps, and improves the efficiency and user satisfaction of the charging network.
Smart Images

Figure CN118966416A_ABST
Abstract
Description
Technical Field
[0001] The patent of this invention relates to a method for site selection and capacity determination of electric vehicle charging stations based on an improved Gray Wolf optimizer, which is of great significance in modern urban planning. Background Art
[0002] The rapid development of electric vehicles has promoted the construction of a clean energy transportation system, and the establishment of an effective charging infrastructure is the core of achieving this goal. The location and capacity of charging stations are important factors affecting the efficiency and service quality of the charging network, which directly affects the acceptability and convenience of electric vehicles. In recent years, various methods have been proposed to solve the problem of optimizing the allocation of electric vehicle charging stations. Traditional solutions usually include several types, such as linear programming, integer programming, and multi-criteria decision-making methods. These methods attempt to balance cost, accessibility, and other factors such as grid capacity and environmental impact. Scholars use mathematical modeling and optimization techniques to determine the location and capacity of charging stations to meet overall demand. However, these methods have high time complexity in solving the problem of location and capacity of charging stations and are difficult to find the optimal location of charging stations in high-dimensional space, resulting in unevenness and inconvenience in the charging network when providing services. With the development of heuristic algorithms, researchers have adopted node-based problem models, reallocation algorithms, queuing theory and service satisfaction, real data integration and other methods to address the unevenness and inconvenience of user charging. These studies are user-centric and integrate various aspects of electric vehicle use into the complexity of the site selection and sizing model to improve user satisfaction and the efficiency of the charging network. The Grey Wolf Optimization Algorithm (GWO) is a typical heuristic algorithm that has been widely used in many studies to solve the site selection and sizing problem of electric vehicle charging stations.
[0003] Since traditional optimization methods are often inefficient in solving the problem of charging station location and capacity determination, this paper is committed to proposing an improved gray wolf algorithm to solve this challenge. In view of the problem that the traditional gray wolf optimization algorithm is prone to fall into local optimality and slow convergence, we propose an electric vehicle charging station location and capacity determination method based on the improved gray wolf optimizer (IEMGWO). The algorithm adopts multiple strategies in combination, and significantly improves the performance of the gray wolf optimization algorithm (GWO) in terms of search efficiency and convergence speed by combining diversified search mechanisms with innovative update rules. First, based on the basic gray wolf optimization algorithm, a new strategy of nonlinear adjustment of the convergence factor a is introduced, which helps to jump out of the local optimal solution and improve the ability of global search; then, the present invention proposes a new position update strategy, which combines three unique methods. Each method introduces a weight formula with a different allocation mechanism to allow individuals to learn from the global optimal position and the individual optimal position at the same time. After that, the position of the gray wolf individual is further updated through these methods to generate three candidate solutions, and the fitness value of each solution is calculated to select the optimal solution to update the position of the individual. Experimental results show that the minimum total charging distance from users to charging stations obtained by the IEMGWO algorithm is better than other algorithms, and compared with other swarm intelligence optimization algorithms, it has better convergence speed and solution accuracy. Summary of the invention
[0004] Although various methods based on the gray wolf optimization algorithm have achieved some success in solving the problem of charging station location and capacity determination, most solutions based on the gray wolf optimization algorithm are prone to falling into local optimal solutions and slow convergence. The existing methods face two major challenges: (1) Traditional methods are not time-efficient in solving the problem of charging station location and capacity determination. The minimum total distance charging station problem usually selects the location of the charging station in two-dimensional or three-dimensional space, which is relatively easy to handle. However, as the problem scale increases, current methods find it difficult to find the optimal location and optimal capacity of the charging station within a reasonable time. Therefore, we introduce a gray wolf optimization algorithm that integrates multiple strategies to solve this problem. The traditional gray wolf optimization algorithm has low search efficiency in high-dimensional space and is difficult to find the optimal solution within a reasonable time. Therefore, an appropriate convergence criterion is required to ensure that the algorithm obtains the optimal solution within a reasonable time. (2) The traditional gray wolf optimization algorithm has a local optimal solution trap when solving the problem of charging station location and capacity determination. The goal of the minimum total distance charging station location and capacity determination problem is to find the optimal charging station location to minimize the total distance for users to reach their respective charging stations. Due to the complexity of the problem, there are a large number of possible local optimal solutions, which causes the Gray Wolf Optimization Algorithm to fall into them and be unable to find the global optimal solution. Especially when there are a large number of charging stations, the area is complex, or the user distribution is uneven, the algorithm is more susceptible to the impact of local optimal solutions and it is difficult to obtain a global optimal solution. In order to solve the above problems, the present invention considers the actual needs of users and the charging distance, and designs a minimum total distance charging station site selection and sizing method based on the improved enhanced multi-strategy Gray Wolf Optimization Algorithm (IEMGWO). The present invention mainly includes six parts: (1) Modeling the charging station site selection and sizing problem. (2) Introducing a nonlinear convergence factor. (3) Introducing a weighted position update formula. (4) A training framework based on the improved Gray Wolf Optimizer. (5) Experimental verification of the effectiveness of the algorithm.
[0005] The following are the contents of the above six parts:
[0006] 1. Model the problem of charging station location and capacity determination. Define the problem based on the problem instance, and define the decision variables, constraints and objective function of the problem of charging station location and capacity determination.
[0007] 2. Introducing a nonlinear convergence factor. The nonlinear convergence factor can provide a large change in the early stage of the algorithm, allowing the algorithm to explore the solution space more extensively, thereby increasing the possibility of finding the global optimal solution. In this invention, the nonlinear convergence factor is introduced into the gray wolf optimization algorithm to improve the algorithm's search efficiency and convergence speed.
[0008] 3. Introduce a weighted position update formula. By assigning different weights to the individual optimal and global optimal positions, the algorithm can maintain the exploration of the global optimal solution while also taking into account the individual historical optimal position, thereby enhancing the global search capability. Through three unique weight allocation strategies, the algorithm can more effectively update the position of the gray wolf individual, generate multiple candidate solutions, and select the optimal solution through fitness evaluation, thereby improving the quality of the solution.
[0009] 4. Training framework based on the improved gray wolf optimizer. First, the specific location of the user demand point and the corresponding charging demand are analyzed and obtained in detail. Next, the number of charging stations that need to be built within a certain range is determined, and the initial values of various parameters used in the model are set. The next step of the model is to initialize the position of the gray wolf group. This step involves simulating the behavior of the gray wolf group to find the optimal location of the charging station. Each gray wolf individual represents a potential charging station layout plan. Finally, the objective function of the charging station site selection and sizing problem is used to calculate the target value of each gray wolf individual to find the optimal charging station location.
[0010] 5. Experimental verification of the effectiveness of the model. Experiments have shown that compared with existing charging station site selection and sizing methods, the present invention has higher quality solutions under similar calculation time, and has higher solution quality and shorter calculation time than most traditional heuristic methods. More importantly, the model is not only highly applicable and practical for the problem of charging station site selection and sizing, but also has strong generalization capabilities and can be applied to site selection decisions for a variety of other large-scale problems.
[0011] The detailed implementation steps taken by the present invention to achieve the above-mentioned purpose are as follows:
[0012] Step 1: Model the location and capacity determination problem of charging stations; Assume that the location set of user demand points I = {x1, x2, x3, ..., x n} and the demand set of all demand points D = {d1, d2, d3, ..., d n}, and comprehensively consider various factors to determine the total number of charging stations to be built as P. Under the premise of a certain number of charging stations, select charging stations with appropriate locations and appropriate capacities in this area to meet the charging needs of electric vehicle users (each demand point is assigned to the charging station closest to it), so that the total distance from all user demand points to their respective charging stations is minimized. The decision variables are defined as follows:
[0013]
[0014] Among them, X i The value of represents whether a charging station is built near the i-th demand point. If so, the value is 1, otherwise it is 0.ij 1 means the jth charging station serves the ith demand point, otherwise 0. The total number of charging stations built is P. A demand point can only be assigned to one charging station, which must be built. Therefore, the following constraints are met:
[0015]
[0016] Among them, a demand point can only be assigned to one charging station, so for any demand point i, its Y ij The sum of is always 1. The capacity of each charging station is calculated using the following formula:
[0017]
[0018] Among them, d i represents the demand of the ith demand point, Y ij 1 means the jth charging station serves the ith demand point, otherwise it is 0. i ,p j ) represents the distance from the i-th demand point to the charging station j. In summary, the objective function of the charging station location and capacity problem can be expressed as:
[0019]
[0020] Step 2: Introduce a nonlinear convergence factor; the nonlinear convergence factor can dynamically adjust the search step size according to the iterative process of the algorithm, so that the algorithm can perform a more refined search when approaching the optimal solution, thereby improving the convergence accuracy. In order to enhance the global and local search capabilities of the algorithm, a nonlinear convergence factor is introduced in the iterative process. As the number of iterations increases, it decreases from 2 and gradually approaches 0. The calculation formula is as follows:
[0021] a(t+1)=max(0.01,2e -0.1a(t) ×(1+0.3cos(a(t))))
[0022] Where t is the current iteration number. When t=1, the value of a is 2. The nonlinear convergence factor is introduced to adjust the adaptive weight coefficient A to enhance the global search ability and local search ability of the algorithm. The calculation formula is as follows:
[0023] A=2a(t)·ra(t)
[0024] Among them, t is the current iteration number, r is a random number between [0,1], |A|≥1 means that the gray wolf performs a global search, and |A|<1 means that the gray wolf performs a local search. a is a nonlinear convergence factor, as shown in the following formula:
[0025] Step 3: Introduce a weighted position update formula; the gray wolf optimization algorithm is prone to fall into local extreme values when solving complex optimization problems, that is, premature convergence. To solve the above problem, the present invention improves the position update formula, uses three different calculation methods to generate three candidate solutions, and then calculates the fitness values of the three candidate solutions respectively, selects the one with the best fitness value as the final solution, and thereby determines the final position of the ω individual.
[0026] The positions of other gray wolves are updated according to the positions of α wolf, β wolf and δ wolf. The update formula is as follows:
[0027] X1=X α -A1·|C1·X α -X|
[0028] X2=X β -A2·|C2·X β -X|
[0029] X3=X δ -A3·|C3·X δ -X|
[0030] Among them, X1, X2 and X3 represent the influence of α-layer wolf pack, β-layer wolf pack and δ-layer wolf pack respectively, and X α , X β and X δ Represent the positions of α wolf, β wolf and δ wolf respectively, X is the current position of the gray wolf individual, A1, A2 and A3 are adaptive weight coefficients, C1, C2 and C3 are random vectors, and the calculation formula is as follows:
[0031] C=2r
[0032] Wherein, r is a random number between [0,1]. The present invention uses three different calculation methods to generate three candidate solutions, and then calculates the fitness values of the three candidate solutions respectively, and selects the one with the best fitness value as the final solution, thereby determining the final position of the ω individual.
[0033] The first candidate solution is to let the individual learn from both the global optimal position and the individual historical optimal position, and introduce a weight formula. The calculation formula is as follows:
[0034]
[0035] D=(1-w)·(r3·(X α -X(t))+(1-r3)·(X1-X(t)))
[0036] Where t is the current iteration number, r3 is a random number between [0,1], and w represents the inertia weight, which decreases linearly from the initial value w1 to the final value w2, calculated as follows:
[0037] w=w1-(w1-w2)×r4+0.1×r5
[0038] The second candidate solution is to assign different weights to X1, X2 and X3 respectively, so as to guide the wolf pack to evolve towards the optimal solution. A dynamic weight strategy is introduced, and the proportional weight is calculated using the improved step length Euclidean distance. r4 and r5 are random numbers between [0,1], and the calculation formula is as follows:
[0039]
[0040] Where t is the current iteration number, T is the total number of iterations, and the weights of the three parts are calculated using the following formula, where ε∈[0,0.01]:
[0041]
[0042] The third candidate solution is to randomly select an individual from the population and perform a differential search with the current individual. At the same time, the memory formula of the individual is improved so that it can remember the optimal solution in its own evolution process. r1, r2 and r3 are random numbers between [0,1], and the calculation formula is as follows:
[0043]
[0044] Substitute the three candidate solutions into the objective function for calculation, and select the one with the best fitness value as the final position of the ω individual. The calculation formula is as follows:
[0045]
[0046] Step 4: Based on the training framework of the improved gray wolf optimizer; in the problem of charging station location selection and sizing, the different location combinations of all charging stations can be regarded as multiple different planning schemes. Therefore, the present invention adopts the gray wolf optimization algorithm that integrates multiple strategies to enhance the global search capability and convergence speed of the traditional gray wolf optimization algorithm. Each gray wolf represents a different charging station placement scheme. The specific steps of using the improved gray wolf optimizer to solve the problem of charging station location selection and sizing are as follows:
[0047] Step 4.1: Obtain the specific location of the user demand point and the corresponding charging demand, and determine the number of charging stations to be built;
[0048] Step 4.2: Population initialization; First, a certain number of gray wolf individuals are randomly generated as the initial population. Each gray wolf individual can be represented as a possible charging station location and its corresponding capacity configuration scheme. The positions and states of these individuals will be continuously updated according to the value of the objective function to gradually find the optimal solution;
[0049] Step 4.3: Objective function definition; The problem of charging pile location and capacity determination is to find the optimal location for all charging stations in a certain area and allocate appropriate capacity so that the sum of the distances from all user demand points to their respective charging stations is minimized while meeting user needs. This problem is converted into a fitness function, which can be expressed as x i represents the location of the i-th demand point, p j represents the jth charging station, dist(x i ,p j ) represents the distance from the i-th demand point to the charging station j, Y ij 1 means that the jth charging station serves the ith demand point, otherwise it is 0;
[0050] Step 4.4: Position update; the positions of other gray wolf individuals are updated according to the positions of wolf α, wolf β and wolf δ. The position update formulas of three different strategies are combined for calculation. The three candidate solutions are substituted into the objective function for calculation. The one with the best fitness value is selected as the final position of individual ω. The calculation formula is as follows:
[0051]
[0052] Among them, F represents the objective function, X1(t+1), X2(t+1) and X3(t+1) represent the three candidate solutions generated, and finally X p (t+1) is the final position of the individual gray wolf, and also represents the planning scheme of a charging station.
[0053] Step 4.5: Update α wolf, β wolf and δ wolf; according to the fitness value, select the three gray wolf individuals with the best fitness value in the current population as α wolf, β wolf and δ wolf respectively. α wolf represents the best solution in the current population, β wolf and δ wolf are the second best and third best solutions respectively.
[0054] Step 4.6: Determine the termination condition; determine whether the maximum number of iterations has been reached or the fitness value is no longer significantly improved. If so, output the current optimal solution, that is, the optimal charging station location plan; otherwise, return to step 4.3 and continue iterating.
[0055] Step 4.7: Determine the capacity of the charging station; when the locations of all charging stations are determined, the nearest charging station is assigned to each demand point, and the capacity of each charging station is determined as the sum of the demand of the demand points served. The calculation formula is as follows:
[0056]
[0057] Among them, d i represents the demand of the ith demand point, Y ij1 means the jth charging station serves the ith demand point, otherwise it is 0. i ,p j ) represents the distance from the i-th demand point to the charging station j.
[0058] Step 4.8: Output the results; output the final charging station site selection and capacity configuration plan, and draw a distribution connection diagram of demand points and charging stations.
[0059] The present invention considers the actual needs of users and charging distance, and proposes an improved gray wolf algorithm to solve the problem of charging station location and capacity determination, and optimizes the total charging distance of users in the charging network, which is called the minimum total distance charging station location and capacity determination method based on the improved enhanced multi-strategy gray wolf optimization algorithm (IEMGWO). The IEMGWO algorithm uses multiple strategies to optimize the original gray wolf algorithm, improving the search strategy and convergence performance of the algorithm. On the one hand, we propose a new strategy of nonlinear adjustment of the convergence factor, which can adjust the convergence speed of the algorithm, so that the algorithm approaches the global optimal solution faster, and obtains a better charging station location and capacity determination scheme within a limited number of iterations, which can greatly shorten the algorithm solution time and improve the efficiency of the algorithm. On the other hand, we propose a new position update strategy, which integrates three different weight allocation formulas to achieve the algorithm individual learning from the global optimal and individual optimal positions at the same time, and then updates the individual position by merging the three methods, and generates three candidate solutions, and then selects the best solution as the position update based on the fitness value, which enhances the global search ability of the algorithm and effectively avoids the local optimal trap, so that it can find the global optimal solution. The algorithm in this paper significantly improves the global search efficiency of the algorithm by combining different search strategies and update strategies, effectively avoiding falling into the local optimal solution, and thus exploring the global optimal solution in a wider search space. The introduction of multiple strategies enables the Grey Wolf Optimization Algorithm to find a better charging station location and capacity solution, providing an efficient tool for solving optimization problems. BRIEF DESCRIPTION OF THE DRAWINGS
[0060] Figure 1 This is a schematic diagram of the charging station site selection and capacity determination problem in the present invention.
[0061] Figure 2 This is the framework diagram of site selection and capacity determination of electric vehicle charging stations based on the improved gray wolf optimizer in the present invention.
[0062] Figure 3 This is a graph showing the variation of the nonlinear and linear convergence factors in the present invention.
[0063] Figure 4 This is the training total distance convergence curve of 30 demand points and 4 charging stations in the present invention.
[0064] Figure 5This is a comparison of the total distance convergence curves of IEMGWO and other algorithms in the present invention.
[0065] Figure 6 This is the distribution connection diagram of 30 demand points and 4 charging stations in the present invention
[0066] Figure 7 This is the Voronoi diagram of 30 demand points and 4 charging stations in the present invention. Figure 8 This is the effectiveness verification convergence curve of different improved modules in the present invention DETAILED DESCRIPTION
[0068] The present invention is further described below with reference to the accompanying drawings and examples.
[0069] The present invention designs an electric vehicle charging station site selection and capacity determination framework based on an improved optimizer for the charging station site selection and capacity determination problem. The convergence rule of the algorithm is optimized by introducing a nonlinear convergence factor, thereby accelerating the convergence speed. In order to overcome the limitation that the traditional algorithm is prone to fall into the local optimum, a new position update rule is adopted, and the position update formula is improved in combination with the weight adjustment mechanism. Three different position update methods are further developed, each of which aims to adjust the search behavior of individual gray wolves in a unique way, generate three candidate solutions, and then select the optimal solution to update the position of the gray wolf by evaluating the fitness of these candidate solutions. These comprehensive strategies significantly enhance the algorithm's ability in global search and avoiding local optimal traps, greatly improving the overall efficiency and accuracy of solving complex site selection problems.
[0070] Step 1: In a charging station site selection and capacity determination problem, 30 demand point locations and corresponding demand quantities are determined within a certain range, and 4 charging stations need to be built. First, randomly generate instances. The coordinates of the 30 demand nodes are (0.03, 0.82), (0.99, 0.64), (0.93, 0.97), (0.77, 0.58), (0.54, 0.10), (0.01, 0.54), (0.42, 0.38), (0.74, 0.94), (0.09, 0.48), (0.24, 0.22), (0.40, 0.90), (0.29, 0.51), (0.31, 0.64), (0.44, 0.86), (0.83, 0.16), (0.56, 0.62), (0.48, 0.68), (0.73, 0.30), (0.11, 0.47), (0.28, 0.05), (0.60, 0.68) ,(0.71,0.45),(0.46,0.68),(0.80,0.31),(0.37,0.01),(0.48,0.50),(0.06,0.80),(0.69,0.20),(0.17,0.71),(0.73,0.59), the demand of 30 demand nodes is 0.49, 0.38, 0.22 , 0.34, 0.93, 0.77, 0.61, 0.82, 0.56, 0.11, 0.45, 0.54, 0.99, 0.49, 0.82, 0.53, 0.19, 0.34, 0.60, 0.71, 0.36, 0.93, 0.19, 0.22, 0.39, 0.49, 0.94, 0.83, 0.24, 0.29. The decision variables defined in the present invention are as follows:
[0071]
[0072] Based on this, under the condition of satisfying the constraints, the objective function of the charging station location and capacity problem is:
[0073]
[0074] Among them, Y ij 1 means the jth charging station serves the ith demand point, otherwise it is 0. i ,p j ) represents the distance from the i-th demand point to the charging station j.
[0075] Step 2: In the present invention, a nonlinear convergence factor is introduced into the Gray Wolf Optimization Algorithm to improve the performance of the algorithm in terms of search efficiency and convergence speed. The linear convergence factor can dynamically adjust the search step size according to the iterative process of the algorithm, so that the algorithm can perform a more refined search when approaching the optimal solution, thereby improving the convergence accuracy. In order to enhance the global search capability and local search capability of the algorithm, a nonlinear convergence factor is introduced in the iterative process, and it decreases from 2 to 0 as the number of iterations increases. The iterative change curves of the linear convergence factor and the nonlinear convergence factor are shown in FIG. Figure 3 As shown in the figure, it can be seen that the value of the nonlinear parameter fluctuates strongly in the early stage, which is conducive to the algorithm to jump out of the local optimal solution and enhance the ability of global search. The value of the nonlinear parameter is smaller in the middle and late stages, which is conducive to long-term local development.
[0076] Step 3: The present invention introduces a weighted position update formula, uses three different calculation methods to generate three candidate solutions, and then calculates the fitness values of the three candidate solutions respectively, and selects the one with the best fitness value as the final solution, so as to determine the final position of the ω individual.
[0077] Step 4: In the problem of charging station location selection and capacity determination, the different location combinations of all charging stations can be regarded as multiple different planning schemes. Therefore, the present invention adopts the gray wolf optimization algorithm that integrates multiple strategies to enhance the global search capability and convergence speed of the traditional gray wolf optimization algorithm. Each gray wolf represents a different charging station placement scheme, and the improved gray wolf optimizer is used to solve the problem of charging station location selection and capacity determination. The specific training process is as follows:
[0078] Step 4.1: First, obtain the demand points and demand quantities of randomly generated instances.
[0079] Step 4.2: Initialize the positions of 30 individual gray wolves, each of which represents a possible charging station location and capacity configuration plan. Four charging stations need to be built, and the location of each charging station is two-dimensional, so each gray wolf position should be a vector similar to ([0.43, 0.73], [0.49, 0.91], [0.79, 0.09], [0.44, 0.18]). Each pair of coordinates represents the location of a charging station. The positions of 30 gray wolves are randomly generated in turn. In this example, the training is set to 100 times.
[0080] Step 4.3: The goal of the charging station site selection and capacity determination in the present invention is to minimize the total distance, and the fitness function is According to the fitness function, the top three wolves with the best fitness values are calculated as α wolf, β wolf and δ wolf.
[0081] Step 4.4: After determining the first three wolves, the positions of other gray wolves in the wolf pack are updated according to the position update formula of the present invention, so that the solved charging station position is closer to the optimal solution.
[0082] Step 4.5: After the wolf pack positions are updated, the positions of α wolf, β wolf and δ wolf are updated according to the value of the fitness function.
[0083] Step 4.6: Determine whether the algorithm has reached the maximum number of iterations of 100, or the calculated fitness value no longer changes significantly. At this time, output the corresponding gray wolf position, which is the best location of the charging station found. In this example, when the algorithm reaches the maximum number of iterations, the minimum total distance found is 5.65, and the location of the optimal charging station is (0.34, 0.06), (0.33, 0.64), (0.77, 0.77), (0.76, 0.32). The total distance change convergence curve during training is as follows Figure 4 As shown in Figure 2, the IEMGWO algorithm has converged at the 10th iteration, and compared with the total distance convergence curves of other algorithms on this instance, Figure 5 As shown, it can be seen that the algorithm has a faster convergence speed and higher convergence accuracy than other algorithms, which verifies the effectiveness of the algorithm.
[0084] Step 4.7: The previous step has found the optimal placement of charging stations. Then, the nearest charging station is assigned to each demand point, and the capacity of each charging station is calculated. In this example, the capacities of the four charging stations are determined to be 1.53kw, 8.06kw, 3.25kw, and 3.09kw respectively.
[0085] Step 4.8: Finally, the location of the optimal charging station and its corresponding capacity are output. The distribution relationship between demand points and charging stations is as follows: Figure 6 As shown, Figure 7 Use the Voronoi diagram to divide the scope of charging stations.
[0086] Assume that the location coordinates and demand of user demand points are randomly generated. A total of 30 demand points and corresponding demand are generated. The number of charging stations to be built is set to 4. The experiment is run 100 times. The number of wolf packs is set to 30, and each wolf represents a planning scheme for a charging station. In order to adapt to different scales, training is performed on the location and capacity of charging stations with 30 and 100 demand points. In training, all instances are randomly generated, and the generated instances are randomly sampled according to a uniform distribution. The coordinates of the user demand points are sampled in the unit area [0,1]×[0,1], and the demand is sampled between [0,1].
[0087] In the verification, the IEMGWO algorithm was first applied to the example of the present invention. Then, GWO, IGWO, MGWO, EGWO, CGWO and NGWO were also applied to the site selection and capacity determination of charging stations as a comparative experiment. Finally, the following variants of the IEMGWO algorithm were performed: (1) IEMGWO-1: In the traditional grey wolf optimization algorithm, only the convergence factor adjustment strategy with nonlinear changes was introduced, and the original position update formula remained unchanged. (2) IEMGWO-2: In the traditional grey wolf algorithm, only the improved position update equation of the present invention was introduced, and three different update rules were used to change the position of the grey wolf without changing the original convergence factor update formula. The variant algorithm and the IEMGWO algorithm were compared on this example as an ablation experiment. The experimental results are shown in the figure. Figure 8 As shown in the figure. The effectiveness of the proposed method is verified by comparative experiments and ablation experiments. The final results show that the IEMGWO algorithm has significantly improved the optimization performance, and it surpasses the existing improved gray wolf optimization algorithm and other swarm intelligence algorithms in terms of overall performance.
Claims
1. A method for site selection and capacity determination of electric vehicle charging stations based on an improved gray wolf optimizer, characterized in that The steps include: Definition: IEMGWO stands for Improved Enhanced Multi-strategy Grey Wolf Optimizer. This invention adopts multiple strategies and significantly improves the performance of the Grey Wolf Optimizer (GWO) in terms of search efficiency and convergence speed by combining diversified search mechanisms with innovative update rules. First, based on the basic Grey Wolf Optimization Algorithm, a new strategy of nonlinear adjustment of the convergence factor a is introduced, which helps to jump out of the local optimal solution and improve the ability of global search; then, a new position update strategy is proposed in this invention, which combines three unique methods. Each method allows individuals to learn from the global optimal position and the individual optimal position at the same time by introducing a weight formula with a different allocation mechanism. After that, the position of the individual Grey Wolf is further updated through these methods to generate three candidate solutions, and the fitness value of each solution is calculated to select the optimal solution to update the position of the individual, so as to find the best position of all charging stations. Finally, the capacity of each charging station is determined by the sum of the demand of all demand points served. Step 1: Model the location and capacity determination problem of charging stations; Assume that the location set of user demand points I = {x1, x2, x3, ..., x n } and the demand set of all demand points D = {d1, d2, d3, ..., d n }, and comprehensively consider various factors to determine the total number of charging stations to be built as P. Under the premise of a certain number of charging stations, select charging stations with appropriate locations and appropriate capacities in this area to meet the charging needs of electric vehicle users (each demand point is assigned to the charging station closest to it), so that the total distance from all user demand points to their respective charging stations is minimized. The decision variables are defined as follows: Among them, X i The value of represents whether a charging station is built near the i-th demand point. If so, the value is 1, otherwise it is 0. ij 1 means the jth charging station serves the ith demand point, otherwise 0. The total number of charging stations built is P. A demand point can only be assigned to one charging station, which must be built. Therefore, the following constraints are met: Among them, a demand point can only be assigned to one charging station, so for any demand point i, its Y ij The sum of is always 1. The capacity of each charging station is calculated using the following formula: Among them, d i represents the demand of the ith demand point, Y ij 1 means the jth charging station serves the ith demand point, otherwise it is 0. i ,p j ) represents the distance from the i-th demand point to the charging station j. In summary, the objective function of the charging station location and capacity problem can be expressed as: Step 2: Introduce a nonlinear convergence factor; the nonlinear convergence factor can dynamically adjust the search step size according to the iterative process of the algorithm, so that the algorithm can perform a more refined search when approaching the optimal solution, thereby improving the convergence accuracy. In order to enhance the global and local search capabilities of the algorithm, a nonlinear convergence factor is introduced in the iterative process. As the number of iterations increases, it decreases from 2 and gradually approaches 0. The calculation formula is as follows: a(t+1)=max(0.01,2e -0.1a(t) ×(1+0.3cos(a(t)))) Where t is the current iteration number. When t=1, the value of a is 2. The nonlinear convergence factor is introduced to adjust the adaptive weight coefficient A to enhance the global search ability and local search ability of the algorithm. The calculation formula is as follows: A=2a(t)·ra(t) Among them, t is the current iteration number, r is a random number between [0,1], |A|≥1 means that the gray wolf performs a global search, and |A|<1 means that the gray wolf performs a local search. a is a nonlinear convergence factor, as shown in the following formula: Step 3: Introduce a weighted position update formula; the gray wolf optimization algorithm is prone to fall into local extreme values when solving complex optimization problems, that is, premature convergence. To solve the above problem, the present invention improves the position update formula, uses three different calculation methods to generate three candidate solutions, and then calculates the fitness values of the three candidate solutions respectively, selects the one with the best fitness value as the final solution, and thereby determines the final position of the ω individual. The positions of other gray wolves are updated according to the positions of α wolf, β wolf and δ wolf. The update formula is as follows: X1=X α -A1·|C1·X α -X| X2=X β -A2·|C2·X β -X| X3=X δ -A3·|C3·X δ -X| Among them, X1, X2 and X3 represent the influence of α-layer wolf pack, β-layer wolf pack and δ-layer wolf pack respectively, and X α , X β and X δ Represent the positions of α wolf, β wolf and δ wolf respectively, X is the current position of the gray wolf individual, A1, A2 and A3 are adaptive weight coefficients, C1, C2 and C3 are random vectors, and the calculation formula is as follows: C=2r Wherein, r is a random number between [0,1]. The present invention uses three different calculation methods to generate three candidate solutions, and then calculates the fitness values of the three candidate solutions respectively, and selects the one with the best fitness value as the final solution, thereby determining the final position of the ω individual. The first candidate solution is to let the individual learn from both the global optimal position and the individual historical optimal position, and introduce a weight formula. The calculation formula is as follows: D=(1-w)·(r3·(X α -X(t))+(1-r3)·(X1-X(t))) Where t is the current iteration number, r3 is a random number between [0,1], and w represents the inertia weight, which decreases linearly from the initial value w1 to the final value w2, calculated as follows: w=w1-(w1-w2)×r4+0.1×r5 The second candidate solution is to assign different weights to X1, X2 and X3 respectively, so as to guide the wolf pack to evolve towards the optimal solution. A dynamic weight strategy is introduced, and the proportional weight is calculated using the improved step length Euclidean distance. r4 and r5 are random numbers between [0,1], and the calculation formula is as follows: Where t is the current iteration number, T is the total number of iterations, and the weights of the three parts are calculated using the following formula, where ε∈[0,0.01]: The third candidate solution is to randomly select an individual from the population and perform a differential search with the current individual. At the same time, the memory formula of the individual is improved so that it can remember the optimal solution in its own evolution process. r1, r2 and r3 are random numbers between [0,1], and the calculation formula is as follows: Substitute the three candidate solutions into the objective function for calculation, and select the one with the best fitness value as the final position of the ω individual. The calculation formula is as follows: Step 4: Based on the training framework of the improved gray wolf optimizer; in the problem of charging station location selection and sizing, the different location combinations of all charging stations can be regarded as multiple different planning schemes. Therefore, the present invention adopts the gray wolf optimization algorithm that integrates multiple strategies to enhance the global search capability and convergence speed of the traditional gray wolf optimization algorithm. Each gray wolf represents a different charging station placement scheme. The specific steps of using the improved gray wolf optimizer to solve the problem of charging station location selection and sizing are as follows: Step 4.1: Obtain the specific location of the user demand point and the corresponding charging demand, and determine the number of charging stations to be built; Step 4.2: Population initialization; First, a certain number of gray wolf individuals are randomly generated as the initial population. Each gray wolf individual can be represented as a possible charging station location and its corresponding capacity configuration scheme. The positions and states of these individuals will be continuously updated according to the value of the objective function to gradually find the optimal solution; Step 4.3: Objective function definition; The problem of charging pile location and capacity determination is to find the optimal location for all charging stations in a certain area and allocate appropriate capacity so that the sum of the distances from all user demand points to their respective charging stations is minimized while meeting user needs. This problem is converted into a fitness function, which can be expressed as x i represents the location of the i-th demand point, p j represents the jth charging station, dist(x i ,p j ) represents the distance from the i-th demand point to the charging station j, Y ij 1 means that the jth charging station serves the ith demand point, otherwise it is 0; Step 4.4: Position update; the positions of other gray wolf individuals are updated according to the positions of wolf α, wolf β and wolf δ. The position update formulas of three different strategies are combined for calculation. The three candidate solutions are substituted into the objective function for calculation. The one with the best fitness value is selected as the final position of individual ω. The calculation formula is as follows: Among them, F represents the objective function, X1(t+1), X2(t+1) and X3(t+1) represent the three candidate solutions generated, and finally X p (t+1) is the final position of the individual gray wolf, and also represents the planning scheme of a charging station. Step 4.5: Update α wolf, β wolf and δ wolf; according to the fitness value, select the three gray wolf individuals with the best fitness value in the current population as α wolf, β wolf and δ wolf respectively. α wolf represents the best solution in the current population, β wolf and δ wolf are the second best and third best solutions respectively. Step 4.6: Determine the termination condition; determine whether the maximum number of iterations has been reached or the fitness value is no longer significantly improved. If so, output the current optimal solution, that is, the optimal charging station location plan; otherwise, return to step 4.3 and continue iterating. Step 4.7: Determine the capacity of the charging station; when the locations of all charging stations are determined, the nearest charging station is assigned to each demand point, and the capacity of each charging station is determined as the sum of the demand of the demand points served. The calculation formula is as follows: Among them, d i represents the demand of the ith demand point, Y ij 1 means the jth charging station serves the ith demand point, otherwise it is 0. i ,p j ) represents the distance from the i-th demand point to the charging station j. Step 4.8: Output the results; output the final charging station site selection and capacity configuration plan, and draw a distribution connection diagram of demand points and charging stations.
Citation Information
Cited By
Comprehensive energy system operation optimization method based on multi-strategy combination optimization
CN119312987A
