A method for planning an electric taxi charging network

Abstract

The application discloses a kind of electric taxi charging network planning method, it is related to charging network planning field;The electric taxi of the present application is the research object, comprehensively considers charging station annual fixed investment and vehicle charging time behavior cost factor, establishes the electric vehicle charging station site selection model with the minimum annual total cost as target, and uses genetic algorithm to optimize site selection.This technical scheme fully considers the total time cost of charging consumption, the fixed investment cost and operation cost of charging station, adds the time cost from charging station to the first search passenger node position after charging in the total time consumption cost of charging behavior, so that the cost calculation is more in line with the actual, and the deviation of actual operation cost is small, the model obtained is more scientific, and the charging network planning obtained is more reasonable, compared with prior art, the scientific station planning of charging station can effectively reduce user time cost, and unnecessary investment on charging station is reduced.

Description

Technical Field

[0001] This invention relates to the field of charging network planning, and more particularly to a method for planning a charging network for electric taxis. Background Technology

[0002] Based on current planning and development needs and the current situation, the number of electric vehicles will inevitably enter a phase of rapid growth. When a large number of electric vehicles are charged in a concentrated manner, the power grid will be significantly affected by this charging load. At the same time, the location of charging stations is also related to charging efficiency and economic issues.

[0003] Related research has used the Monte Carlo method to analyze the correlation between electric vehicle charging demand and travel time, deriving a probabilistic model of charging demand. Based on this, corresponding electric vehicle charging station site selection can be carried out. Domestic and international researchers have optimized charging station site selection from two aspects: actual data extraction and model prediction. Some studies have extracted and processed actual taxi GPS data, analyzed relevant required parts to obtain charging demand with spatiotemporal attributes, and established corresponding spatiotemporal demand models to discuss the optimal charging station site selection. Some scholars have proposed a site selection model considering traffic flow, aiming to maximize the target traffic flow through the selected site, constructing a traffic capture site selection model. Other research, based on the variable weight Voronoi diagram and using a hybrid discrete particle swarm optimization algorithm, optimizes the search for the optimal charging station site by satisfying the maximum electric vehicle charging demand and charging station service radius within the region, with the goal of minimizing the target annual comprehensive cost, thus improving the accuracy of site selection. However, due to the high complexity of the variable weight Voronoi diagram algorithm, combining it with the hybrid discrete particle swarm optimization algorithm can easily get trapped in local optima and fail to converge to the optimal solution. Meanwhile, the algorithm simultaneously calculates the number and location of charging stations. It seeks the optimal solution for all schemes with different numbers of charging stations and then compares the results, leading to a long solution time. Regarding the model, this study considers the time cost for electric taxis to travel to charging stations and the cost of waiting for charging, but does not consider the time cost for them to reach the first passenger search node after charging. This results in a bias in cost calculation, deviating from actual operating costs and preventing the charging network planning from reaching its optimal state. Therefore, the existing research requires further improvement in both algorithm efficiency and model. Summary of the Invention

[0004] The technical problem to be solved and the technical task proposed by this invention is to improve and refine existing technical solutions, and to provide a method for planning electric taxi charging networks to achieve the optimal spatial layout of charging stations. To this end, this invention adopts the following technical solution.

[0005] A method for planning a charging network for electric taxis includes the following steps:

[0006] 1) Obtain information, including information on the planned charging station area, electric taxis, and charging station parameters;

[0007] 2) Obtain the electric taxi charging network planning model. The electric taxi charging network planning model aims to minimize the total annual cost of electric vehicle charging stations, and comprehensively considers the total time consumption cost of charging behavior, the annual fixed investment cost of charging stations, and the annual operation and maintenance cost of charging stations. The total time consumption cost of charging behavior includes three parts: the time cost for the electric taxi to travel from its current location to the charging station, the time cost for waiting and charging at the charging station, and the time cost for the electric taxi to reach the first search passenger node location after charging is completed.

[0008] 3) Calculate the feasible range and optimal solution for the number of charging stations to be built. Calculate the range for the number of charging stations to be built based on the total charging power demand of electric vehicles in the region. Calculate the optimal solution for the number of charging stations to be built based on the electric taxi charging network planning model, and obtain the optimal number of charging stations.

[0009] 4) Use a genetic algorithm to calculate the specific location of the charging station;

[0010] Using the electric taxi charging network planning model function as the fitness function of the genetic algorithm, the coordinates of random charging stations are initialized based on the determined optimal number of charging stations. The location information is then encoded and used as the solution variable to obtain the optimal spatial location of the charging station, thus forming an electric taxi charging network planning scheme.

[0011] This invention decouples the optimal number and optimal distribution of charging stations, solving the problem step-by-step, which improves the algorithm's convergence and time efficiency. From a social benefit perspective, an electric vehicle charging station site selection model is established, fully considering the total time cost of charging, the fixed investment cost of charging stations, and the operation and maintenance cost. The time cost of reaching the first search passenger node after charging is added to the overall time cost of the charging activity, making the cost calculation more realistic and less deviating from actual operating costs. This results in a more scientific model and a more reasonable charging network plan.

[0012] This technical solution uses a genetic algorithm for solving the problem. It can flexibly handle various complex constraints, has strong optimization capabilities, and good convergence ability.

[0013] As a preferred technical means: In step 1), the charging station planning area information includes road length and intersection coordinates; electric taxi parameters include vehicle model, battery capacity and average driving speed; charging station parameters include the number of chargers, power and simultaneous rate; the information obtained also includes: the empty driving time cost of taxi drivers and taxi order information, the taxi order information includes origin information and destination information.

[0014] As a preferred technical approach: In step 2), the electric taxi charging network planning model function is:

[0015] min P=T+C c +C v (1)

[0016] Where T represents the total time cost of the charging activity; C c C represents the annual fixed investment cost of a charging station; v This represents the annual operation and maintenance cost of a charging station.

[0017] As a preferred technical means: the total time cost T of the charging process is:

[0018] T = ∑ i∈A P ik t ik N i +∑ k∈M (∑ i∈A P ik t k1 N i )+(∑ j∈A (∑ i∈A P ik N i )P kj t kj ) (3)

[0020] Where A is the set of traffic nodes, i,j∈A; M is the set of charging station location nodes; N i P represents the number of taxis traveling from node i to the charging station node. ik Let t be the probability of charging from traffic node i to charging station location k; ik t represents the shortest travel time from traffic node i to charging station location k; k1 P represents the waiting and charging time of electric vehicles at charging stations. kj Let t be the probability of traveling from charging station location k to traffic node j; kj This represents the shortest travel time from charging station location k to traffic node j. The overall time consumption cost of the above charging behavior is comprehensively considered and can realistically simulate actual conditions.

[0021] As a preferred technical approach: The annual fixed investment cost of a charging station includes investments in charging facilities, auxiliary equipment, and land purchase and leasing. The annual fixed investment cost of a charging station is:

[0022]

[0023] The number of distribution transformers configured in the charging station is n. iThe unit cost of the configured distribution transformer is 'a', and the number of charging piles in the charging station is 'm'. i The unit price of a charging pile is b, and the cost of basic investment in the charging station is c. i The discount rate is r0, and the design operating life is z;

[0024] The annual operation and maintenance cost of a charging station includes employee salaries, equipment maintenance costs, and equipment depreciation costs. This cost is expressed using a proportional factor η. The annual operation and maintenance cost of a charging station is:

[0025]

[0026] As a preferred technical approach: In step 3), the strategy for selecting the number of charging stations is to plan based on the total charging power demand of electric vehicles in the region; where the total number of electric vehicles N... EV Given the rated battery capacity W of the selected vehicle model, the total charging demand for electric vehicles in the region can be calculated; combined with the maximum capacity limit S of the charging station. max With minimum capacity limit S min The range of the number of charging stations can be obtained as shown in equation (2).

[0027]

[0028] Where, N min N represents the minimum planned number of electric vehicle charging stations in the region. max N represents the maximum planned number of electric vehicle charging stations in the region. min ≤N≤N max .

[0029] As a preferred technical means: In step 4), the fitness function value of the initial scheme is calculated based on the coordinates of the initialized random charging stations. The next generation population is generated by randomly transforming the position codes through selection, crossover, and mutation operations. The selection operation inherits individuals with low fitness functions to the next generation, that is, the charging station distribution scheme with low overall cost is retained to continue to participate in the optimization. The crossover operation partially swaps the position codes of two relatively better individuals after the selection operation. The newly generated individual inherits the advantages of the original individual, that is, the newly generated charging station layout scheme may combine the advantages of the two schemes. The mutation operation randomly mutates the position codes of each relatively better individual, so that the new individual seeks optimization in the vicinity of the existing individuals. That is, each relatively better charging station layout scheme is fine-tuned and optimized in random directions around itself. The fitness function of the new population is calculated to evaluate whether the charging station distribution scheme represented by the position code is optimized. When the iteration condition is met, the optimal charging station scheme is output. Otherwise, the selection, crossover, and mutation operations of the generated population position codes are repeated until the iteration condition is met.

[0030] Beneficial Effects: This invention decouples the optimal number and optimal distribution of charging stations, solving the problem step-by-step, which improves the convergence and time efficiency of the algorithm. From a social benefit perspective, an electric vehicle charging station site selection model is established, fully considering the total time cost of charging, the fixed investment cost of charging stations, and the operation and maintenance cost. The time cost of reaching the first search passenger node after charging is added to the overall time cost of the charging behavior, making the cost calculation more realistic and less deviating from actual operating costs. The resulting model is more scientific, and the resulting charging network planning is more reasonable.

[0031] This technical solution uses a genetic algorithm for solving the problem. It can flexibly handle various complex constraints, has strong optimization capabilities, and has good convergence capabilities.

[0032] Time cost is a crucial factor affecting taxi profitability. This technical solution comprehensively considers taxi time costs along with the construction and operation costs of charging stations, utilizing a genetic algorithm to ultimately derive a planning scheme. Scientific site planning for electric vehicle charging stations effectively reduces user time costs, avoids excessive and unnecessary investment in charging stations, and improves charging station utilization. Lowering the operating costs of electric taxis ensures their income, contributing to the healthy development of the electric taxi industry, increasing the market share of electric taxis, and helping to reduce exhaust emissions.

[0033] This technical solution uses probabilistic modeling, which can meet the needs of large-scale planning problems with high efficiency and low computational cost. It solves the computational efficiency problem in complex road networks and with multiple charging stations. Compared with methods that directly accumulate the time costs of each individual vehicle, the method proposed in this patent is more applicable. It avoids the problem of excessive computation when using simple accumulation on a 365-day year scale. Attached Figure Description

[0034] Figure 1 This is a flowchart of the present invention.

[0035] Figure 2 This is the road network map of the planning area of ​​this invention.

[0036] Figure 3 This is a graph showing the relationship between the total annual cost of the charging station and the number of charging stations according to the present invention.

[0037] Figure 4 This is the optimized charging station site selection diagram of the present invention. Detailed Implementation

[0038] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings.

[0039] Due to the unique operating methods of taxis, time costs are a crucial factor affecting their profitability. The longer the passenger-carrying time of an electric taxi, the shorter the time spent traveling to a charging station and then returning to find passengers, resulting in higher profits. This invention aims to minimize the total annual cost of electric vehicle charging stations, comprehensively considering both the annual fixed investment in charging stations and the time costs associated with vehicle charging. The specific process is as follows... Figure 1 As shown, the electric taxi charging network planning method includes the following steps:

[0040] Step 1: Obtain information such as road network information, electric taxi and charging station parameters for the planned area.

[0041] Input basic road network information for the planning area, including road length and intersection coordinates. Input electric vehicle model, battery capacity, and average driving speed. Input the number, power, and simultaneous charging rate of charging stations. Input parameters such as the empty driving time cost for taxi drivers.

[0042] Step 2: Obtain the electric taxi charging network planning model.

[0043] This model aims to minimize the total annual cost of electric vehicle charging stations, taking into account both the annual fixed investment in charging stations and the behavioral costs of vehicle charging time. The function is shown in equation (1).

[0044] min P=T+C c +C v (1)

[0045] Where T represents the total time cost of the charging activity; C c C represents the annual fixed investment cost of a charging station; v This represents the annual operation and maintenance cost of a charging station.

[0046] The strategy for selecting the number of charging stations is based on the total charging power demand of electric vehicles in the region. Given the total number of electric vehicles N... EV Given the rated battery capacity W of the selected vehicle model, the total charging demand for electric vehicles in the area can be calculated. This is combined with the maximum capacity limit S of the charging stations. max With minimum capacity limit S min The range of the number of charging stations can be obtained as shown in equation (2).

[0047]

[0048] Where, N min N represents the minimum planned number of electric vehicle charging stations in the region. max N represents the maximum planned number of electric vehicle charging stations in the region. min ≤N≤N max .

[0049] The total time cost of charging consists of three parts: the time cost of traveling from the current location to the charging station, the time cost of waiting and charging at the charging station, and the time cost of traveling from the charging station to the first search passenger node after charging. The objective function is to measure the sum of the time cost of traveling to the charging station, the time cost of waiting and charging at the charging station, and the time cost of traveling from the charging station to the first search passenger node after charging, as shown in equation (3).

[0050] T = ∑ i∈A P ik t ik N i +∑ k∈M (∑ i∈A P ik t k1 N i )+(∑ j∈A (∑ i∈A P ik N i )P kj t kj ) (3)

[0052] Where A is the set of traffic nodes, i,j∈A; M is the set of charging station location nodes; N i P represents the number of taxis traveling from node i to the charging station node. ik Let t be the probability of charging from traffic node i to charging station location k; ik t represents the shortest travel time from traffic node i to charging station location k; k1 P represents the waiting and charging time of electric vehicles at charging stations. kj Let t be the probability of traveling from charging station location k to traffic node j; kj Let k be the shortest travel time from charging station location k to traffic node j.

[0053] The fixed costs of a charging station consist of two parts: the annual fixed investment cost and the annual operation and maintenance cost. The annual fixed investment cost mainly covers investments in charging facilities, auxiliary equipment, land purchase and leasing, etc., while the annual operation and maintenance cost consists of employee salaries, equipment maintenance expenses, etc.

[0054] Annual fixed investment cost of charging stations:

[0055]

[0056] The number of distribution transformers configured in the charging station is n. i The unit cost of the configured distribution transformer is 'a', and the number of charging piles in the charging station is 'm'. i The unit price of a charging pile is b, and the cost of basic investment in the charging station is c.i The discount rate is r0, and the design operating life is z.

[0057] Operation and maintenance costs include employee wages, equipment maintenance fees, equipment depreciation fees, etc. The factors affecting these costs fluctuate greatly, and it is usually impossible to find a unified pattern for accurate mathematical expression. Therefore, a proportional factor η can be set to express the annual operation and maintenance costs, as shown in equation (5).

[0058]

[0059] Step 3: Calculate the feasible range and optimal solution for the number of charging stations to be built.

[0060] The minimum number of planned charging stations N can be calculated from equation (2). min And the maximum number of charging stations N max The optimal number of charging stations is determined based on the total annual cost of the charging stations.

[0061] Step 4: Use a genetic algorithm to determine the exact location of the charging station.

[0062] Using formula (1) as the fitness function of the genetic algorithm, and after determining the optimal number of charging stations in step 3, the coordinates of the random charging stations are initialized, and the location information is encoded and used as the solution variable. The optimal spatial location of the charging stations is obtained through the genetic algorithm, forming a charging network scheme. Specifically, the fitness function value of the initial scheme is calculated based on the initial coordinates of the random charging stations. The next generation of population is generated by randomly transforming the location codes through selection, crossover, and mutation operations. The selection operation inherits individuals with low fitness functions to the next generation, that is, the charging station distribution scheme with low overall cost is retained to continue to participate in the optimization. The crossover operation partially exchanges the location codes of two individuals that are relatively better after the selection operation. The newly generated individual inherits the advantages of the original individual, that is, a newly generated charging station layout scheme may combine the advantages of the two schemes. The mutation is to randomly mutate the location codes of each relatively better individual, so that the new individual can optimize in the vicinity of the existing individuals, that is, to fine-tune and optimize each relatively better charging station layout scheme in the random direction around itself. The fitness function of the new population is used to evaluate whether the charging station distribution scheme represented by the location code is optimized. When the iteration condition is met, the optimal charging station scheme is output. Otherwise, the selection, crossover and mutation operations of the generated population location codes are repeated until the iteration condition is met.

[0063] This technical solution establishes a quantifiable economic model for electric taxi charging network planning and uses a genetic algorithm to solve for the optimal spatial layout of charging stations. This effectively reduces user time costs while avoiding excessive and unnecessary investment in charging stations, thus improving charging station utilization.

[0064] The following are further explanations using specific examples:

[0065] Step 1: Input information such as road network information, electric taxi and charging station parameters for the planned area.

[0066] This paper analyzes a plot of land in a certain region of Zhejiang Province and plans the construction of electric vehicle charging stations within this area. The road network in this area is as follows: Figure 2 As shown, the area covers 62.5 square kilometers and has 72 road network nodes.

[0067] The electric taxi model selected is the BYD E6, which has a battery capacity of 60 kWh, an average vehicle speed of 30 km / h, a power of 120 kW per charger in the charging station, a maximum of 30 chargers and a minimum of 10 chargers installed in the charging station, a charging simultaneous rate of 0.9, an empty driving time cost of 14.32 yuan / hour for taxi drivers, a genetic algorithm iteration count of 100 times, a discount rate of 0.08, and a designed operating life of 20 years.

[0068] Step 2: Obtain the electric taxi charging network planning model.

[0069] The electric taxi charging network planning model is established based on equations (1) to (5).

[0070] Step 3: Calculate the feasible range and optimal solution for the number of charging stations to be built.

[0071] The minimum number of charging stations N can be calculated from equation (2). min =4, maximum number of charging stations N max =12. Therefore, we have... Figure 3 The graph showing the relationship between the total annual cost of a charging station and the number of charging stations indicates that the total annual cost is lowest when the planned number of charging stations is 9.

[0072] Step 4: Use a genetic algorithm to determine the exact location of the charging station.

[0073] Nine coordinates were selected as the initial location coordinates for the electric vehicle charging station. Then, a genetic algorithm was used to optimize the location of these coordinates. The optimized charging station location coordinates are shown below. Figure 4As shown in the diagram. Specifically, an initial population is generated, where each individual is encoded by nine random coordinates to form a gene. The fitness function of the initial scheme is calculated based on the coordinates of the initialized random charging stations. The next generation of the population is produced by randomly transforming the position codes through selection, crossover, and mutation operations, optimizing in random directions. Selection involves passing individuals with low fitness functions to the next generation, meaning that charging station distribution schemes with low overall cost are retained for further optimization. Crossover involves partially exchanging the position codes of two relatively superior individuals after selection, with the newly generated individual inheriting the advantages of the original individual; that is, a newly generated charging station layout scheme may combine the advantages of both schemes. Mutation involves randomly mutating the position codes of relatively superior individuals, allowing the new individual to optimize in the vicinity of existing individuals, essentially fine-tuning each relatively superior charging station layout scheme in random directions around itself. The fitness function of the new population is calculated to evaluate whether the charging station distribution scheme represented by the position codes is optimized. When the iteration condition is met, the optimal charging station scheme is output; otherwise, the selection, crossover, and mutation operations on the generated population position codes are repeated until the iteration condition is met.

[0074] from Figure 4 As shown in the site selection data, the charging stations are distributed relatively evenly, without any areas being excessively dense or sparse. The charging stations tend to be located in areas with high vehicle density, which increases service coverage and extends the service range. This is because the algorithm model comprehensively considers the time cost of traveling from the current location to the charging station, the time cost of waiting and charging at the charging station, and the time cost of returning to the first search passenger node after charging. This ensures that the charging station distribution considers both proximity to areas with high demand for electric taxi charging and ensuring that the overall charging distance for taxis is not too far, effectively avoiding unnecessary waste of resources.

[0075] The electric taxi charging network planning method shown above is a specific embodiment of the present invention, which has demonstrated the substantial features and progress of the present invention. According to actual use needs, equivalent modifications in shape, structure, etc. can be made to it under the guidance of the present invention, all of which are within the protection scope of this solution.

Claims

1. A method for planning an electric taxi charging network, characterized in that The method comprises the following steps: 1) obtaining information, including charging station planning area information, electric taxi and charging station parameters; 2) obtaining an electric taxi charging network planning model, the electric taxi charging network planning model taking the minimum annual total cost of electric vehicle charging stations as the target, and comprehensively considering the overall time consumption cost of charging behavior, annual fixed investment cost of charging stations and annual operation and maintenance cost of charging stations; wherein the overall time consumption cost of charging behavior includes three parts, namely, the time cost of electric taxis going to charging stations from the current location, the time cost of waiting and charging in charging stations, and the time cost of electric taxis going to the first passenger node location from charging stations after charging; the function of the electric taxi charging network planning model is: min P = T + C c + C v (1) wherein, T represents the total time consumption cost of charging behavior; C c represents the annual fixed investment cost of the charging station; C v represents the annual operation and maintenance cost of the charging station; 3) calculating the feasible range and optimal solution of the number of charging stations, calculating the range of the number of charging stations based on the total demand for electric vehicle charging power in the region, and sequentially calculating the annual total cost corresponding to the number of each candidate charging station in the range based on the electric taxi charging network planning model, and determining the optimal number of charging stations that minimizes the annual total cost by comparison; 4) calculating the specific location of the charging station using a genetic algorithm; Taking the function of the electric taxi charging network planning model as the fitness function of the genetic algorithm, initializing the coordinates of the random charging station according to the determined optimal number of charging stations, encoding the position information as a solution variable to obtain the optimal charging station space location and form an electric taxi charging network planning scheme; calculating the fitness function value of the initial scheme according to the coordinates of the initialized random charging station, randomly transforming the position code through selection, crossover and mutation operations, and optimizing the next generation population in the random direction; the fitness function of the new population is calculated to evaluate whether the charging station distribution scheme represented by the position code is optimized, and the optimal charging station scheme is output when the iteration condition is met, otherwise the cycle of selection, crossover and mutation operations on the generated population position code is continued until the iteration condition is met.

2. The method of claim 1, wherein: In step 1), the charging station planning area information includes road length and intersection coordinate information; the electric taxi parameters include vehicle model, battery capacity and average driving speed; The charging station parameters include the number, power and simultaneous rate of charging machines; the obtained information further includes the empty driving time cost of taxi drivers and taxi order information, and the taxi order information includes starting point information and ending point information.

3. The method of claim 2, wherein: Charging behavior total time consumption cost T To: T = ∑ i∈A P ik t ik N i + ∑ k∈M (∑ i∈A P ik t k1 N i ) + (∑ j∈A (∑ i∈A P ik N i ) P kj t kj ) (3) wherein, is a set of traffic nodes, ; is a set of charging station location nodes; is a node from which the number of taxis to the charging station node; is a probability of charging from the traffic node to the charging station location is a shortest travel time from the traffic node to the charging station location is a waiting and charging time for an electric vehicle at the charging station; is a probability from the charging station location k to the traffic node j; is a shortest travel time from the charging station location k to the traffic node j.

4. The electric taxi charging network planning method according to claim 3, characterized in that: The annual fixed investment cost of charging stations includes investment for charging facilities, auxiliary equipment, land purchase and leasing, and the annual fixed investment cost of charging stations is: (4) Wherein, the number of distribution transformers configured by the charging station is , the unit cost of the configured distribution transformer is , the number of charging piles in the charging station is , the unit price of the charging pile is , the cost for the basic investment of the charging station is , the discount rate is , and the design operation life is ; The annual operation and maintenance cost of the charging station includes staff salary, equipment maintenance cost, and equipment depreciation cost. The annual operation and maintenance cost of the charging station is expressed by setting a proportional factor The annual operation and maintenance cost of the charging station is: (5)。 5. The method of claim 4, wherein: In step 3), the strategy of selecting the number of charging stations is to plan according to the total demand of electric vehicle charging power in the region; in the total number of electric vehicles N EV , the rated capacity of the battery of the selected vehicle model W The total demand of electric vehicle charging in the region can be calculated under the known condition; combined with the maximum capacity limit of the charging station S max and the minimum capacity limit S min , the range of the number of charging stations can be obtained as shown in formula (2) (2) wherein, N min the minimum number of planned charging stations for electric vehicles in the region, N max the maximum number of planned charging stations for electric vehicles in the region, the number of planned charging stations for electric vehicles in the region N min ≤ N ≤ N max .

6. The method of claim 5, wherein: In step 4), the selection operation is to genetically inherit the individuals with low fitness function to the next generation, i.e. to keep the charging station distribution scheme with low comprehensive cost to continue participating in the optimization; the crossover operation is to partially exchange the position codes of two individuals relatively optimal after the selection operation, and the newly generated individual inherits the advantages of the original individual, i.e. the newly generated charging station layout scheme may integrate the advantages of the two schemes; the mutation is to randomly mutate the position codes of each individual relatively optimal, so that the new individual optimizes in the vicinity of the existing individual, i.e. to slightly adjust and optimize the individual relatively optimal in the random direction around the individual.

Images