A low-carbon public charging facility site selection method

Abstract

The application discloses a low-carbon public charging facility site selection method, and the method steps comprise the following steps: S1, a plurality of populations are randomly generated, and optimal individuals in a plurality of different populations are obtained; S2, a plurality of different populations are subjected to game through a game model, and game results and a game base number of the plurality of different populations are obtained; S3, if the game base number does not converge, the population adjusts a diversity related value of the population according to the game result, performs a crossover operation and a mutation operation, updates the optimal individual, and step S2 is performed; if the game base number converges, a global optimal individual is obtained, and then a charging facility layout is obtained; through the method, the charging facility layout is reasonable, and carbon emission of the charging facility site selection in a charging process is reduced.

Description

Technical Field

[0001] This invention relates to a method for site selection of low-carbon public charging facilities, and the technical field is the field of charging station technology. Background Technology

[0002] Incorporating carbon emission considerations into the planning and layout of charging infrastructure has become a core issue in the global energy transition and transportation electrification process. With the rapid growth of electric vehicle ownership, the coupling relationship between charging demand and power system load is becoming increasingly close. Traditional charging network planning models often focus on single dimensions such as user convenience, grid capacity, and construction costs, neglecting the potential impact of the dynamic relationship between charging behavior and energy structure and spatiotemporal distribution on carbon emissions. Research shows that the carbon emission reduction benefits of electric vehicles throughout their entire life cycle are highly dependent on the cleanliness of the electricity source for charging. Furthermore, the spatiotemporal aggregation effect of charging loads may force local power grids to utilize high-carbon-emission peak-shaving units during specific periods, increasing the carbon footprint of electric vehicles by more than 30% compared to theoretical values. Especially in energy systems dominated by thermal power, disorderly charging layout may trigger a vicious cycle of "increased electricity demand - fossil fuel consumption - rising carbon emissions," creating a structural contradiction with the goals.

[0003] In existing technologies, carbon emissions are rarely considered in the layout of charging facilities. In one feasible implementation example of this embodiment, a planning method for electric vehicle charging stations based on the CRITIC method and non-cooperative game theory is disclosed on the China Patent Network, with application number 202111221924.0. However, the method in this patent does not consider how to reduce carbon emissions in the layout of charging facilities, which will increase the carbon emissions or carbon factor of electric vehicles during the charging process. Summary of the Invention

[0004] The purpose of this invention is to reduce the carbon emissions of electric vehicles during the charging process by further optimizing the layout of charging facilities, while ensuring that the layout of charging facilities is reasonable.

[0005] Another objective of this invention is to reduce the probability of the obtained charging facility layout getting trapped in a local optimum by combining a game theory model with an improved genetic algorithm.

[0006] To achieve the above objectives, this invention provides a method for site selection of low-carbon public charging facilities. The method includes the following steps: S1, randomly generating several populations and obtaining the optimal individual from several different populations; S2, having several different populations play a game through a game model to obtain the game results and game cardinality of several different populations; S3, if the game cardinality does not converge, adjusting the diversity-related values ​​of the populations according to the game results, performing crossover and mutation operations, updating the optimal individual, and executing step S2; if the game cardinality converges, obtaining the globally optimal individual, and thus obtaining the charging facility layout.

[0007] Preferably, step S1 includes: the element generation probability distribution of the individual is positively correlated with the weighted carbon flow rate of the adjacent power system nodes at the element corresponding position of the individual.

[0008] Preferably, step S1 includes: the distance between the adjacent power system nodes and the individual is the first few minimum distances.

[0009] Preferably, step S1 includes: the degree of the adjacent power system nodes of the individual is not greater than the degree-related threshold.

[0010] Preferably, step S3 includes: in the mutation operation, the element mutation probability distribution of the individual is related to the diversity correlation value of the population to which the individual belongs; if the diversity correlation value of the population is lower, then the probability of element mutation is higher in the element mutation probability distribution.

[0011] Preferably, step S1 includes: the optimal charging facility layout corresponding to the individual, which has the highest fitness among all the charging facility layouts corresponding to the individuals in the population.

[0012] Preferably, step S3 includes: when performing the crossover operation between the individuals, the probability of selecting the individual is positively correlated with the fitness of the individual.

[0013] Preferably, step S3 includes: in the crossover operation between the individuals, selecting a crossover point such that the absolute difference between the difference values ​​between the sub-individuals is maximized.

[0014] Preferably, step S2 includes: in the game model, the game result includes: the relative score of the individual; the relative score of the population is related to the carbon emission related index, supply and demand matching index and quantity related index of the charging facility layout corresponding to the population.

[0015] Preferably, step S2 includes: if the relative ranking of the correlation function of the population decreases, then the relative score of the population decreases.

[0016] The beneficial effects of this invention are that, while ensuring a reasonable layout of charging facilities, it reduces carbon emissions or carbon dioxide emissions from electric vehicles during the charging process by optimizing the layout of these facilities. Furthermore, the combination of a game theory model and a genetic algorithm reduces the probability of the global optimum getting trapped in a local optimum. Attached Figure Description

[0017] Figure 1 This invention relates to a flowchart of one implementation of the method.

[0018] Figure 2 This is a type of regional division map.

[0019] Figure 3 This is a diagram illustrating a cross-operation between individuals.

[0020] Figure 4 This invention relates to a flowchart of one implementation of the method.

[0021] Figure 5 This invention relates to a flowchart of one implementation of the method.

[0022] Figure 6 This is a system architecture diagram. Detailed Implementation

[0023] Example 1: This example discloses a specific implementation flow of the method involved in the present invention. (See attached document.) Figure 1 .

[0024] In this embodiment, the specific implementation of the method is described in a feasible implementation example of this embodiment below.

[0025] S1. Obtain the topology information of the power system in the current area. Based on the topology information of the power system in the current area and the distribution of charging demand, determine all locations of charging facilities. On this basis, randomly generate several populations and obtain the best individual from several different populations.

[0026] It should be noted that the current distribution of charging demand in the region can be obtained through statistics, which will not be elaborated here; the layout of charging facilities is equivalent to the site selection of charging facilities.

[0027] The current area is divided into several different sub-regions. In different sub-regions, the charging demand may vary due to differences in region type or region indicators.

[0028] The impact of regional type and indicator differences on electric vehicle charging demand is rooted in the coupling mechanism between spatial functional differentiation and user behavior patterns. Commercial areas, due to high-density commuting, logistics delivery, and shared mobility vehicle aggregation, exhibit pulse-like bursts of daytime charging demand, with charging periods highly overlapping with peak commercial activities, and the proportion of fast charging is significantly higher than in other areas. In contrast, residential areas are dominated by slow charging at night, with the charging load curve driven by residents' work and rest patterns, showing a typical pattern of gradual increase during off-peak hours at night.

[0029] The charging demand of each sub-region is processed by: obtaining the average charging demand of each sub-region; and dividing the sum of the charging demands of each sub-region and its neighboring sub-regions by the sum of the charging demands of all sub-regions to obtain the weight value of the charging demand of each sub-region.

[0030] By using weighted threshold values, it is determined whether each sub-region is a potential location for charging facilities.

[0031] For a given sub-region, if the weight value of the sub-region is greater than the weight value threshold, then the sub-region is a possible location for a charging station; if the weight value of the sub-region is less than the weight threshold, then the sub-region is not a possible location for a charging station.

[0032] After determining all locations for charging facilities, the average carbon flow rate of adjacent power system nodes for all locations is obtained, thereby obtaining the weighted average carbon flow rate of adjacent power system nodes for all locations.

[0033] For a given location, the power system nodes that meet the following conditions are considered to be adjacent power system nodes of that location.

[0034] Condition 1: The degree of a node in the power system is less than the degree-related threshold.

[0035] In one feasible implementation example of this embodiment, there is a certain location where the power system has nodes A, B and C, and the degree correlation threshold is 4. If the degrees of nodes A, B and C are 5, 4 and 3 respectively, then nodes A and B satisfy condition one.

[0036] Condition 2: The power system nodes are the top K nodes that are closest to the selected location.

[0037] In one feasible implementation example of this embodiment, there exists a certain location. The power system has nodes A, B, C, D, and E, and the value of K is 3. The distances between nodes A and D and the location are as follows: the distance between node A and the location is greater than the distance between node B and the location; the distance between node B and the location is greater than the distance between node C and the location; the distance between node C and the location is greater than the distance between node D and the location. Therefore, the distances between nodes A and B satisfy condition two.

[0038] By selecting power system nodes that meet both condition one and condition two, the adjacent power system nodes of each site are determined, and the carbon flow rate of the adjacent power system nodes is obtained.

[0039] Since the carbon flow rate of each node in the power system varies greatly at different times, it is necessary to obtain the average carbon flow rate of adjacent power system nodes. After determining the average carbon flow rate of adjacent nodes, the weighted average carbon flow rate of adjacent power system nodes is obtained.

[0040] The weighted average carbon flow rate of adjacent power system nodes is as follows.

[0041] For a given location, if the location has m adjacent power system nodes, the weighted average carbon flow rate of the adjacent power system nodes of the location is as follows.

[0042] For the i-th adjacent power system node at this location, the expression for the standard deviation of carbon flow rate of this node during the time period t is as follows.

[0043] Its expression Among them, Cfr i,j Cfr represents the carbon flow rate of the i-th adjacent power system node in the j-th time period. i,avg Let be the average carbon flow rate of the i-th adjacent power system node during the time period t.

[0044] For a given location, the weighted average carbon flow rate of the adjacent power systems at that location is expressed as follows.

[0045] Its expression

[0046] After determining the weighted average carbon flow rate for all charging facility locations, an improved genetic algorithm is executed several times to obtain the optimal individuals from several populations. The execution process is as follows.

[0047] Several individuals are randomly generated. During the generation of individuals, the probability of an individual being generated as an element is related to the weighted average carbon flow rate of the location corresponding to that element.

[0048] That is, the lower the weighted average carbon flow rate of the location corresponding to the element, the higher the probability that the element generates a value indicating that there is a charging facility at the location; the higher the weighted average carbon flow rate of the location corresponding to the element, the lower the probability that the element generates a value indicating that there is a charging facility at the location.

[0049] In this embodiment, the value of an element can be 1 or 0. 1 indicates that a charging facility is provided at the address corresponding to the element, and 0 indicates that no charging facility is provided at the address corresponding to the element.

[0050] If there are n locations in the current region, the expression for the probability distribution of the element generation corresponding to the i-th location is as follows.

[0051] The expression for the element value representing the probability distribution of the location where charging facilities are located is as follows:

[0052] The expression for the element value representing the probability distribution that no charging facility is located at the address corresponding to that element is as follows:

[0053] Among them, WCfr i σ is the weighted average carbon flow rate for the i-th site selection; i,max Let be the standard deviation of the maximum carbon flow rate of the adjacent power system nodes of the i-th location; A indicates that there are charging facilities at the location corresponding to the element; B indicates that there are no charging facilities at the location corresponding to the element.

[0054] After generating a number of individuals, these individuals form a population.

[0055] After executing the above process several times, several populations are obtained. Among these populations, the individual with the highest fitness is selected as the optimal individual for the population.

[0056] The fitness of this individual is specifically the weighted sum of the carbon emission index of the charging facility layout, the charging facility supply and demand matching index, and the charging facility quantity related index. Among them, the carbon emission index has the largest weight value, and the weight of the carbon emission index is not less than 0.5.

[0057] The carbon emissions from the layout of charging facilities are: the sum of the demand load corresponding to the location with an element value of 1 multiplied by the weighted average carbon flow rate of the adjacent power system nodes at the location.

[0058] Its expression Where S1 is the set of elements with a value of 1 for an individual, Load i This represents the charging demand load for location selection.

[0059] The carbon emission index for the layout of charging facilities is: Among them, C max This represents the sum of the maximum carbon emissions for the location corresponding to an element value of 1. Specifically, it is the sum of the maximum load for the location corresponding to an element value of 1 multiplied by the weighted average carbon flow rate of the adjacent power system nodes of the location. Its expression is... Load i,max The maximum load for address i.

[0060] The supply-demand matching index of charging facility layout is: the ratio of the sum of the maximum loads of the locations corresponding to the element value of 1 to the total demand load of the locations corresponding to the element value of 1, with the natural logarithm as the base.

[0061] Its expression Among them, Load i,max The maximum load for address i.

[0062] The quantity-related index of charging facility layout is: the value raised to the power of the number of elements with a value of 1 divided by the negative number of the total number of sites, with the natural logarithm as the base.

[0063] Its expression Where Sd represents the number of elements with a value of 1 in an individual, specifically the number of charging facilities corresponding to that individual; Sd_SUM represents the total number of sites selected.

[0064] The expression for an individual's fitness is f = w C *CAdi+w A *Adi+w S *SAdi, where w C The weighting values ​​corresponding to the carbon emission index of individual charging facility deployment; w A The weight values ​​corresponding to the supply and demand matching index of charging infrastructure deployment; w S The weight values ​​corresponding to the index related to the number of charging facilities deployed.

[0065] It should be noted that, in generating individuals, the addressing load corresponding to an element with a value of 1 is obtained in the following way.

[0066] For a given sub-region (which is not a location for charging facilities), obtain the distance between each charging facility location (where a charging facility is located) and the sub-region, the average historical traffic flow between the sub-region and each charging facility location, the average arrival time between the sub-region and each charging facility location, and the number of roads between the sub-region and each charging facility location.

[0067] Based on the parameters mentioned above between each charging facility location and the sub-region, the flow probability between the sub-region and the charging facility location is obtained, and then the expected traffic flow between the sub-region and the charging location is calculated, and finally the demand load of the charging location is calculated.

[0068] The flow probability between sub-regions and charging facility locations is obtained using a Logit model, the expression of which is described below.

[0069] It should be noted that the flow probability refers to the probability that a trolley will flow to the location of a charging facility when there are trolleys that need charging in the sub-region.

[0070] The expression for the flow probability between the i-th sub-region and the j-th charging facility location is as follows:

[0071] Among them, U i,j The expression U i,j =β1*Dist i,j +β2*Ht i,j +β3*Atoa i,j +β4*Nor i,j Dist i,j Ht is the distance between the i-th sub-region and the j-th charging facility location. i,j Atoa represents the average historical traffic flow from the i-th sub-region to the j-th charging facility site selection. i,j Nor represents the average arrival time of vehicles between the location of the i-th sub-region and the j-th charging facility. i,j The number of roads between the i-th sub-region and the j-th charging facility location.

[0072] For each trolleybus, the probability distribution of its charging demand can be approximated by a normal distribution. Thus, the expected load demand of the trolleybus can be calculated based on the probability distribution of its charging demand. The probability density function expression of the charging demand probability distribution is as follows.

[0073] expression Where s is the charging load demand of the electric vehicle; σ s The standard deviation of the charging demand load for electric vehicles; μ s The average value of the charging load demand for electric vehicles.

[0074] Based on the above expression, the expected charging demand load of the trolley is obtained by multiplying the charging demand load of the trolley by the integral of the probability density function of the charging demand load of the trolley.

[0075] Its expression is Ecn=∫s*f(s)ds.

[0076] When the expected charging load of electric vehicles in the current area is obtained, the load of electric vehicles in the sub-area on the charging site is calculated based on the flow probability between the sub-area and the charging facility site, thereby calculating the load of the charging facility site.

[0077] Specifically, the load of charging facility site selection is equal to the electric vehicle demand load of each sub-region multiplied by the sum of the flow probabilities between the sub-region and the charging facility site selection.

[0078] For the load with the i-th charging address, its expression Load i =∑ j∈Cr Ld j *Ecn*P i,j , among which, Ld j P represents the number of trams in subregion j; i,j Let Cr be the flow probability between the location of the i-th charging facility and sub-region j; Cr is the set of sub-regions in the current region.

[0079] To further understand the above methods for obtaining charging location load demand, an example will be used to illustrate this in detail below.

[0080] See Figure 2 ,exist Figure 2 Among the sub-regions, there are sub-regions 1 to 9. Sub-regions 2, 6 and 7 are where charging stations are located and have charging facilities. In the current region, the expected charging demand load of electric vehicles is 37 kWh.

[0081] The number of trolleybuses requiring charging in sub-areas 1 to 9 are 201, 376, 451, 110, 53, 243, 141, 298, and 322, respectively.

[0082] Taking sub-region 2 as an example, the probability of a vehicle from sub-region 1 flowing into sub-region 2 is 17%; the probability of a vehicle from sub-region 2 flowing into sub-region 2 is 20%; the probability of a vehicle from sub-region 3 flowing into sub-region 2 is 6%; the probability of a vehicle from sub-region 4 flowing into sub-region 2 is 14%; the probability of a vehicle from sub-region 5 flowing into sub-region 2 is 19%; the probability of a vehicle from sub-region 6 flowing into sub-region 2 is 31%; the probability of a vehicle from sub-region 7 flowing into sub-region 2 is 15%; the probability of a vehicle from sub-region 8 flowing into sub-region 2 is 23%; and the probability of a vehicle from sub-region 9 flowing into sub-region 2 is 7%.

[0083] Therefore, the expected load of sub-region 2 is:

[0084] (0.17*201+0.2*376+451*0.06+110*0.14+53*0.19+243*0.31+141*0.15+298*0.23+322*0.09)*37K Wh=13168.3KWh.

[0085] That is, the expected charging load of sub-region 2 is 13168.3 kWh.

[0086] Taking sub-region 6 as an example, the probability of a vehicle from sub-region 1 flowing into sub-region 6 is 27%; the probability of a vehicle from sub-region 2 flowing into sub-region 6 is 42%; the probability of a vehicle from sub-region 3 flowing into sub-region 6 is 51%; the probability of a vehicle from sub-region 4 flowing into sub-region 6 is 67%; the probability of a vehicle from sub-region 5 flowing into sub-region 6 is 62%; the probability of a vehicle from sub-region 6 flowing into sub-region 6 is 39%; the probability of a vehicle from sub-region 7 flowing into sub-region 6 is 57%; the probability of a vehicle from sub-region 8 flowing into sub-region 6 is 46%; and the probability of a vehicle from sub-region 9 flowing into sub-region 6 is 71%.

[0087] Therefore, the expected load of the sub-region is:

[0088] (0.27*201+0.42*376+451*0.51+110*0.67+53*0.62+243*0.39+141*0.57+298*0.46+322*0.7)*37K Wh=40196.06KWh.

[0089] That is, the expected charging load of sub-region 6 is 40196.06 kWh.

[0090] Taking sub-region 7 as an example, the probability of a vehicle from sub-region 1 flowing into sub-region 7 is 56%; the probability of a vehicle from sub-region 2 flowing into sub-region 7 is 38%; the probability of a vehicle from sub-region 3 flowing into sub-region 7 is 43%; the probability of a vehicle from sub-region 4 flowing into sub-region 7 is 19%; the probability of a vehicle from sub-region 5 flowing into sub-region 7 is 19%; the probability of a vehicle from sub-region 6 flowing into sub-region 7 is 30%; the probability of a vehicle from sub-region 7 flowing into sub-region 7 is 28%; the probability of a vehicle from sub-region 8 flowing into sub-region 7 is 31%; and the probability of a vehicle from sub-region 9 flowing into sub-region 7 is 22%.

[0091] Therefore, the expected load of the sub-region is:

[0092] (0.56*201+0.38*376+451*0.43+110*0.19+53*0.19+243*0.3+141*0.28+298*0.31+322*0.22)*37K Wh=27969.78KWh.

[0093] That is, the expected charging load of sub-region 7 is 27969.78 kWh.

[0094] In this way, we can calculate the expected charging demand load of the charging facility site corresponding to the element with an individual element value of 1 in the population, regard the expected charging demand load as the charging demand load of the charging facility site, and then calculate the fitness of all individuals in a population.

[0095] After obtaining the best individuals from several populations, proceed to step S2.

[0096] S2. Several individuals with different optimal outcomes engage in a game through a game model to obtain the game results (relative scores of the populations) and the game base in this round of the game for several different populations.

[0097] The optimal individual in a population and the population's diversity value, to some extent, reflect the potential of the population's solution space. This potential specifically refers to the probability that the optimal individual generated by the population is the global optimal solution.

[0098] If the population has a high diversity correlation value and a high fitness of the best individual in the population, then the probability that the best individual generated in the subsequent generation of the population is the global optimum is higher, which means that the population has greater potential; if the population has a low diversity correlation value and a low fitness of the best individual in the population, then the population has less potential.

[0099] In this embodiment, the diversity correlation of the population is obtained by using the population's genomic entropy, which is described below.

[0100] The gene entropy of a population is specifically the average gene entropy of each species in the population.

[0101] For the i-th gene in the population, the entropy of the gene can be expressed as: H i =Σ v∈Asov -p(v)*ln(p(v)), where p(v) represents the probability of taking the value v; Asov represents the set of values ​​for the gene.

[0102] If there are k genes in a population, the expression for the diversity correlation value of the population is as follows.

[0103] Its expression

[0104] The following example will illustrate this in detail.

[0105] In one feasible implementation example of this embodiment, there are four types of genes in a certain population, with the gene value being either 1 or 0. For the first type of gene, 60% of the individuals in the population have a gene value of 1, and 40% of the individuals have a gene value of 0. For the second type of gene, 37% of the individuals in the population have a gene value of 1, and 63% of the individuals have a gene value of 0. For the third type of individual, 70% of the individuals in the population have a gene value of 1, and 30% of the individuals have a gene value of 0. For the fourth type of individual, 25% of the individuals in the population have a gene value of 1, and 75% of the individuals have a gene value of 0.

[0106] The entropy of the first gene is: -(0.6*ln(0.6)+0.4*ln(0.4)), approximately 0.67 (rounded to two decimal places).

[0107] The entropy of the second gene is: -(0.37*ln(0.37)+0.63*ln(0.63)), approximately 0.66 (rounded to two decimal places).

[0108] The entropy of the third gene is: -(0.7*ln(0.7)+0.3*ln(0.3)), approximately 0.61 (rounded to two decimal places).

[0109] The entropy of the fourth gene is: -(0.25*ln(0.25)+0.75*ln(0.75)), which is approximately 0.56 (rounded to two decimal places).

[0110] In this example, the population diversity correlation value is 0.63 (rounded to two decimal places).

[0111] It should be noted that in this embodiment, since the element can only take two values, the entropy of each gene is no greater than ln2, and therefore the population diversity correlation value is no greater than ln2.

[0112] By observing the game between the best individuals in each population, we can obtain the relative score of a particular population, which represents the game result of the best individual in that population in this round of the game.

[0113] The game process is as follows.

[0114] Before the first round of the game, each population needs to be assigned an initial relative score, which is recorded as the relative score of the previous round of the game, and the game cardinality is generated based on the initial relative score.

[0115] Each population obtains diversity-related values ​​and optimal individual fitness values ​​of other populations, and generates a relative score for each population in this round of the game based on the obtained parameters.

[0116] For a given population, the higher the diversity correlation value and the fitness of the best individual among other populations, the higher the relative score of the population; conversely, the lower the diversity correlation value and the fitness of the best individual among other populations, the lower the relative score of the population.

[0117] For a given population, the relative score for that population in this round of the game is: the relative score of the population in the previous round of the game minus the relative score of the population in the previous round of the game multiplied by the rate of change of the correlation function between the population's diversity and the optimal individual fitness multiplied by the relative ranking of the correlation function in the previous round of the game divided by the relative ranking of the correlation function in this round of the game minus the value 1.

[0118] For the i-th population, in the t-th round of the game, the specific game model can be represented as (the relative score of the population) as:

[0119] The relevant variables in the above formula are as follows.

[0120] Rr i,t This represents the relative score of the i-th population in the t-th round of the game.

[0121] C represents the penalty coefficient, which can be adjusted by those skilled in the art based on the actual situation of the current region; however, it will not be elaborated upon here.

[0122] f i,t This represents the fitness of the best individual in the i-th population during the t-th round of the game.

[0123] Drv i,t This represents the diversity-related value of the i-th population in the t-th round of the game.

[0124] g is the correlation function between the diversity correlation value of the population and the fitness of the best individual in the population, and its expression is...

[0125] rank t (g(f i,t Drv i,t The expression 'rank' represents the relative ranking function of the correlation function of the i-th population among all populations. This relative ranking function is related to the value of the correlation function of the population. If the correlation function of the population ranks higher relative to all populations, the relative ranking function value of the population is lower; if the correlation function of the population ranks lower relative to all populations, the relative ranking function value of the population is larger. t (g)=β*Ps t (g) 2 Pst (g) represents the ranking of the correlation function value of the population among the correlation function values ​​of all populations. In one feasible implementation example of this embodiment, there exists a population whose correlation function value is the largest among all populations. Therefore, the Ps of this population is... t The value of (g) is 1. Similarly, if the correlation function value is the second largest, then the population's Ps... t The value of (g) is 2; β is an adjustment coefficient, and in practical applications, the recommended value range is between 0.61 and 1.38.

[0126] As can be seen from the above relative score values, if a certain population's correlation function value ranks lower among all populations in round t compared to its ranking in round t-1, then the relative score of that population decreases in round t; conversely, if a certain population's correlation function value ranks higher in round t compared to its ranking in round t-1, then the relative score of that population increases in round t.

[0127] In the t-th round of the game, when the relative score of the population decreases, if the correlation function value of the population is large, the degree of decrease in the relative score of the population in the t-th round of the game will be smaller, and the degree of punishment will be smaller; if the correlation function value of the population is small, the degree of decrease in the relative score of the population in the t-th round of the game will be larger, and the degree of punishment will be larger.

[0128] In the t-th round of the game, an increase in the relative score of the population signifies an increase in its potential. However, this potential has an upper limit, and further increases in the relative score become more difficult when the population's correlation function value falls within a certain range. In practical applications, due to factors such as time and construction costs, within a certain range, an increase in the relative score will not lead to a significant improvement in the performance of the charging facility layout corresponding to the optimal individual in the population. Therefore, when the relative score of the population increases, the portion of this increase gradually diminishes as the population's correlation function value increases.

[0129] After the optimal individual in each population has completed the game, the optimal game result for each population (relative score of the population) is obtained, and the game base for this round of game is obtained based on the game result of each population.

[0130] In a certain round of the game, the game is obtained by: the weighted sum of the relative scores of each population with respect to the numerical squares of the correlation function.

[0131] In the t-th round of the game, the cardinality of the game in the t-th round is expressed as follows:

[0132] Where Sz represents the population size.

[0133] S3. Determine whether the game base number has converged based on the first convergence threshold. If the game base number has not converged, the population adjusts the diversity-related value of the population according to the game result, performs several crossover and mutation operations, updates the optimal individual, and executes step S2. If the game base number has converged, the globally optimal individual is obtained, thereby obtaining the charging facility layout corresponding to the globally optimal individual.

[0134] The specific implementation process for this step is as follows.

[0135] After a round of the game ends and the game base for that round is generated, the game base for the previous round is obtained. The absolute difference between the game base for this round and the game base for the previous round is obtained. If the absolute difference is greater than the first convergence threshold, the game base does not converge. Then, each population adjusts its diversity correlation value according to the game result. After adjusting the diversity correlation value, several crossover and mutation operations are performed until the best individual in each population converges. The best individual in each population is updated, and step S2 is executed.

[0136] The implementation process of the above procedure is as follows.

[0137] After the game ends, each population adjusts its population diversity-related value based on the game results. The adjustment process is as follows: for a certain population, if the relative score of the population increases in this round of the game, the population diversity-related value is not adjusted; if the relative score of the population decreases in this round of the game, the population diversity-related value is adjusted. The specific process is as follows.

[0138] During the adjustment process, the target diversity correlation value of the population is related to the game outcome (relative score of the population). For the i-th population, the specific relationship between the game outcome and the target diversity correlation value of that population is as follows.

[0139] Its relational expression Drv i,t,tr =Drv i,t +μ*(ln2-Drv i,t ),in, ε is an adjustment coefficient, which can be adjusted by those skilled in the art according to the actual situation, and will not be elaborated here.

[0140] The adjustment process is as follows: In the population, new individuals are generated according to the element generation probability distribution of each element in the individual, and the population is updated. After updating the population, the best individual in the population is updated. If the diversity correlation value of the updated population is not less than the target diversity correlation value, the above adjustment process ends. If the diversity correlation value of the updated population is not greater than the target diversity correlation value, the adjustment process continues.

[0141] After all populations have been adjusted, several crossover and mutation operations are performed in all populations.

[0142] Specifically, in a certain population, two individuals are randomly selected by roulette. The probability of an individual being selected is the fitness function of the population divided by the sum of the fitness functions of all individuals in the population, as expressed below.

[0143] The probability of the i-th individual in the population being selected is:

[0144] The specific selection method is as follows: generate a numerical range interval based on the probability of an individual being selected in the population, randomly generate two values, and for the l-th individual and the m-th individual, if the two values ​​fall into the numerical range interval corresponding to the l-th individual and the numerical range interval corresponding to the m-th individual respectively, then select the two individuals for crossover operation.

[0145] During the crossover operation, the selected crossover point needs to satisfy the condition that the sum of the differences between the corresponding sub-individuals is maximized, or the crossover point is randomly selected for crossover between individuals.

[0146] In this embodiment, the intersection point between individuals needs to satisfy the condition that the absolute difference in the degree of difference between the corresponding sub-individuals is the largest.

[0147] For sub-individuals a and b, the degree of difference between individuals is specifically as follows:

[0148] Where, when a i With b i When the values ​​are the same, G(a) i ,b i The value of ) is 0; when a i With b i The values ​​are different, G(a) i ,b i The value of ) is 0.

[0149] In one feasible implementation process of this embodiment, see [reference]. Figure 3 ,exist Figure 3 In the given information, there exist individuals 1 and 2. We choose an intersection point P such that the child individuals x1, x2, y1, and y2 satisfy the following conditions: The value is the largest.

[0150] After the crossover operation, the elements in the resulting offspring are mutated.

[0151] The probability distribution of element mutations in offspring is related to the diversity correlation value of the population. Specifically, the lower the diversity correlation value of the population to which the individual belongs, the higher the probability of element mutation for the individual; conversely, the higher the diversity correlation value of the population to which the individual belongs, the lower the probability of mutation for the individual.

[0152] For ease of description, we will set the element mutation event to V and the element no mutation event to NV.

[0153] For the i-th element (also called a gene) in an individual, if the diversity correlation value of the population is Drv, then the probability of an element mutation event occurring for that element is: P i,V =0.02 + 0.03 * e -θ*Drv The probability that no mutation event occurs in elements of this type is P. i,NV =0.98-0.03*e -θ*Drv .

[0154] In one executable implementation example of this embodiment, if the element value is 1 or 0, then when the element mutation event occurs, if the element value is 1, the value of the element after mutation is 0; if the element value is 0, the value of the element after mutation is 1.

[0155] In one executable implementation example of this embodiment, if the element value is A or B, then when the element mutation event occurs, if the element value is A, the value of the mutated element is B; if the element value is B, the value of the mutated element is A.

[0156] After mutation is complete, obtain the two newly generated individuals, and remove the two individuals with the lowest fitness from the population containing the parent and mother individuals. Update the population. After updating the population, update the best individual in the population. If the termination condition is not met, continue to perform crossover and mutation operations in the population. If the termination condition is met, continue to perform crossover and mutation operations in the population.

[0157] In one implementation process of this embodiment, the termination condition is: the number of crossover and mutation operations performed in the population is not less than the set number.

[0158] In one implementation process of this embodiment, the termination condition is: after performing a certain number of crossover and mutation operations, the difference between the fitness of the best individual in the updated population and the fitness of the best individual in the unupdated population is less than the second convergence threshold.

[0159] After the population update and the update of the best individual in each population are completed, step S2 is executed. Each population plays the next round of the game through the game model based on the diversity-related value and the fitness of the best individual in the population.

[0160] If the game's base converges, then the charging facility layout corresponding to the globally optimal individual is obtained.

[0161] The method for obtaining the globally optimal individual is as follows: among the optimal individual sets in several different populations, select the individual with the highest fitness as the globally optimal individual.

[0162] After obtaining the globally optimal individual, the charging facility layout corresponding to the globally optimal individual is obtained, thereby obtaining the final determined charging facility layout.

[0163] The method of this embodiment reduces carbon emissions during the charging process of electric vehicles in areas with complex power systems, while ensuring a reasonable layout and sufficient number of charging facilities. By combining a game theory model and a genetic algorithm, each population can adjust its diversity-related values ​​in real time according to the game results, thereby increasing population diversity. This reduces the probability of the globally optimal individual getting trapped in a local optimum without affecting the solution efficiency. As a result, the method of this invention can provide a reasonable layout of charging facilities in complex power systems.

[0164] Example 2: This example discloses a specific implementation process of the method involved in the present invention, which can be found in the following description. Figure 4 .

[0165] In this embodiment, the implementation of the method of the present invention is described below.

[0166] S1. Randomly generate several different populations. Perform crossover and mutation operations several times in each population until a termination condition is met. After the termination condition is met, obtain the best individual from each population. In this embodiment, the criteria for determining the optimal individual are the same as those for determining the optimal individual in the population involved in Embodiment 1, and will not be repeated here.

[0167] In this embodiment, the specific procedures for crossover and mutation operations between individuals are consistent with those in Embodiment 1, and will not be repeated here.

[0168] S2. Several different populations engage in a game through a game model to obtain the game results of each population and the game cardinality in this round of the game. If the game cardinality does not converge, each population adjusts its diversity-related values ​​according to the game results, updates the population and the best individual in the population, and executes step S1 after the update operation is completed; if the game cardinality converges, execute step S3.

[0169] In this embodiment, the game model involved is the same as that involved in Embodiment 1, and will not be described again here.

[0170] In this embodiment, the game cardinality convergence judgment process is the same as that in Embodiment 1, and will not be described again here.

[0171] In this embodiment, the process for adjusting the diversity-related values ​​of each population is the same as that for adjusting the diversity-related values ​​of each population in Embodiment 1, and will not be repeated here.

[0172] S3. Obtain the individual with the highest fitness in the optimal set of individuals in each population, regard this individual as the global optimal individual, and obtain the charging facility layout corresponding to the global optimal individual.

[0173] The method of this embodiment reduces carbon emissions during the charging process of electric vehicles in areas with relatively simple power systems, while ensuring a reasonable layout of charging facilities and taking into account the number of charging facilities. By combining game theory models and genetic algorithms, each population can adjust its diversity-related values ​​in real time according to the game results, thereby increasing the diversity of the population. This reduces the probability of the globally optimal individual getting trapped in a local optimum without affecting the solution efficiency. As a result, the method of this invention can provide a reasonable layout of charging facilities in complex power systems.

[0174] Example 3: This example discloses a specific implementation process of the method involved in the present invention. See details below. Figure 5 .

[0175] In this embodiment, the specific implementation process of the method is as follows.

[0176] The specific implementation process of step S1 in this embodiment is the same as that in embodiment 1, and will not be repeated here. The specific implementation process of step S2 in this embodiment is the same as that in embodiment 1, and will not be repeated here.

[0177] Alternatively, the specific implementation process of step S1 in this embodiment is the same as that in embodiment 2, and will not be repeated here. The specific implementation process of step S2 in this embodiment is the same as that in embodiment 2, and will not be repeated here.

[0178] In this embodiment, the specific implementation process of step S3 is as follows.

[0179] After obtaining the best individuals from each population, the best individuals from each population are combined into a new population. In the new population, several crossover and mutation operations are performed to update the population and the global best individual.

[0180] In this embodiment, the specific implementation process of step S4 is as follows.

[0181] If the termination condition of the new population is met, the best individual in the new population is the globally optimal individual, and the charging facility layout of the globally optimal individual is obtained; if the termination condition of the new population is not met, then step S3 is executed.

[0182] In one implementation process of this embodiment, the specific process of crossover operation between individuals is as follows: the crossover point is selected randomly, and after the crossover point is selected, the individuals exchange elements (genes) according to the crossover point.

[0183] In one implementation process of this embodiment, the cross-operation between individuals is the same as the specific process of the cross-operation between individuals involved in Embodiment 1 or Embodiment 2.

[0184] In one implementation process of this embodiment, the mutation operation of an individual is a completely random mutation, and the element mutation probability distribution is independent of the diversity correlation value of the population to which the individual belongs, and its mutation probability is 3% to 5%.

[0185] In one implementation process of this embodiment, the specific process of individual mutation operation is the same as that of individual mutation operation involved in Embodiment 1 or Embodiment 2.

[0186] In one implementation of this embodiment, the termination condition for the new population is: the number of crossover and mutation operations performed is not less than a set number.

[0187] In one implementation process of this embodiment, the termination condition of the new population is: after performing a certain number of crossover and mutation operations, the difference between the fitness of the updated global best individual and the fitness of the original global best individual is less than the third convergence threshold.

[0188] In one implementation of this embodiment, the globally optimal individual is the individual with the highest fitness among all individuals in the new population.

[0189] Example 4 discloses an implementation of the methods involved in Examples 1 to 3 on a system.

[0190] The system architecture involved in this embodiment is as follows.

[0191] See Figure 6 The system includes: several execution modules, a central execution module, and a result feedback module.

[0192] The execution module's functions include: obtaining diversity-related values ​​for each population; adjusting the diversity of each population based on the game results; updating the best individual in the population; performing crossover operations between individuals; and performing mutation operations on individual elements.

[0193] The functions of the central execution module include: obtaining the fitness of the best individual in each population and the correlation value of the diversity of each population; generating the game results (relative scores of the population) of each population based on the fitness of the best individual in each population and the diversity of each population; and obtaining the globally best individual.

[0194] The implementation of the method involved in Example 1 on the system involved in this example is as follows.

[0195] The central execution module generates the initial relative scores for each population and generates the game cardinality based on the initial relative scores.

[0196] The execution module generates several populations through random generation, obtains the best individual in each population and its fitness, acquires the diversity-related values ​​of the populations, and transmits them to the central execution module.

[0197] After receiving the fitness and diversity values ​​of the best individual in all populations, the central module generates the game results for each population and the game cardinality in this round of game based on the received data. It then determines whether the game cardinality has converged. If converged, the convergence event information is transmitted to the execution module, which terminates the relevant process and obtains the globally optimal individual, transmitting its data to the result feedback module. If convergence fails, the game results for each population and the non-convergence event are sent to the execution module. The execution module adjusts the diversity values ​​of the populations based on the game results, performing several crossover and mutation operations until the termination condition is met. The updated fitness and diversity values ​​of the best individual in each population are then transmitted to the central execution module, which executes the above process until the game cardinality converges, obtaining the charging facility layout corresponding to the globally optimal individual.

[0198] The implementation of the method involved in Example 2 on the system involved in this example is as follows.

[0199] The execution module generates several populations through random generation, performs several crossover and mutation operations, obtains the best individual and its fitness from the populations, acquires the diversity-related values ​​of the populations, and transmits them to the central execution module.

[0200] After receiving the fitness and diversity-related values ​​of the best individual in all populations, the central module generates the game results for each population and the game cardinality in this round of game based on the received data. It then determines whether the game cardinality has converged. If converged, the convergence event information is transmitted to the execution module, which terminates the relevant process and obtains the globally optimal individual, transmitting its relevant data to the result feedback module. If convergence fails, the game results for each population and the non-convergence event are sent to the execution module. The execution module adjusts the diversity-related values ​​of the populations based on the game results, updates the populations, and transmits the fitness and diversity-related values ​​of the best individual in the updated populations to the central execution module. This process is repeated until the game cardinality converges, and the charging facility layout corresponding to the globally optimal individual is obtained.

[0201] The implementation of the method involved in Example 3 on the system involved in this example is as follows.

[0202] By implementing the above method, the optimal individuals in each population are obtained, forming a population. The population information is then transmitted to the execution mechanism to obtain the charging facility layout corresponding to the globally optimal individual.

[0203] Example 5 discloses an application scenario of the system involved in Example 4.

[0204] When carrying out new energy transformation of the power system, if most of the charging facilities in a certain urban area are already in place, it would be costly to demolish and rebuild the old charging facilities. Moreover, it would be impossible to complete the replacement in a short period of time, and the carbon emissions generated during the construction period cannot be ignored.

[0205] The system described in Example 4 obtains the charging facility layout corresponding to the globally optimal individual. The system described in Example 4 can be deployed on a server. After obtaining the charging facility layout corresponding to the globally optimal individual, the server uploads the information of the charging facility layout to the cloud. The charging facility managers can use the information of the charging layout to ensure that the charging stations located in the charging facility layout are operating normally, and to shut down the charging stations located in other charging facility layouts during certain periods.

[0206] This information will then be transmitted to each vehicle via a server using methods such as broadcasting or system push.

[0207] This invention has illustrated its purpose, technical solutions, and beneficial effects through specific embodiments. However, these embodiments are merely examples to demonstrate the application of the invention and do not constitute a limitation on the scope of protection of this invention. We explicitly state that any reasonable modifications, equivalent substitutions, or technical improvements guided by the spirit and principles of this invention should be included within the scope of protection of this invention. This means that as long as these changes do not deviate from the core idea and basic function of the invention, they should be protected by patent rights. The scope of protection of this invention should be broad, including all directly obvious variations as well as non-obvious innovations that a technical expert can reasonably deduce from the disclosure of this invention. This broad protection aims to promote further research and development based on this invention, while ensuring that its innovativeness and practicality are fully protected by law.

Claims

1. A method for site selection of low-carbon public charging facilities, characterized in that, The steps of the method include: S1. Randomly generate several populations and obtain the best individual from several different populations; S2. Several different populations play a game through a game model to obtain the game results and game cardinality of several different populations; S3. If the game base does not converge, the population adjusts its diversity-related values ​​according to the game result, performs crossover and mutation operations, updates the population-related information, and executes step S2; if the game base converges, the globally optimal individual is obtained, and then the charging facility layout is obtained.

2. The method for site selection of low-carbon public charging facilities according to claim 1, characterized in that, Step S1 includes: the element generation probability distribution of the individual is positively correlated with the weighted carbon flow rate of the adjacent power system nodes at the element corresponding position of the individual.

3. The method for site selection of low-carbon public charging facilities according to claim 2, characterized in that, Step S1 includes: the distance between the adjacent power system nodes and the individual is the first few minimum distances.

4. The method for obtaining the layout of low-carbon public charging facilities according to claim 2, characterized in that, Step S1 includes: the degree of the adjacent power system nodes of the individual is not greater than the degree-related threshold.

5. A method for site selection of low-carbon public charging facilities according to any one of claims 1 to 4, characterized in that, Step S3 includes: in the mutation operation, the element mutation probability distribution of the individual is related to the diversity correlation value of the population to which the individual belongs; if the diversity correlation value of the population is lower, then the probability of element mutation is higher in the element mutation probability distribution.

6. A method for site selection of low-carbon public charging facilities according to any one of claims 1 to 4, characterized in that, Step S1 includes: The optimal charging facility layout for the individual has the highest fitness among all the charging facility layouts for all individuals in the population.

7. The method for site selection of low-carbon public charging facilities according to claim 6, characterized in that, In step S3, the probability of selecting an individual is positively correlated with the fitness of that individual during the crossover operation between the individuals.

8. The method for site selection of low-carbon public charging facilities according to claim 7, characterized in that, Step S3 includes: in the crossover operation between the individuals, selecting a crossover point such that the absolute difference between the difference values ​​between the sub-individuals is maximized.

9. A method for site selection of low-carbon public charging facilities according to any one of claims 1 to 3, characterized in that, Step S2 includes: in the game model, the game result includes: the relative score of the individual; the relative score of the population is related to the carbon emission related index, supply and demand matching index and quantity related index of the charging facility layout corresponding to the population.

10. The method for site selection of low-carbon public charging facilities according to claim 9, characterized in that, Step S2 includes: if the relative ranking of the correlation function of the population decreases, then the relative score of the population decreases.

Images