Electric vehicle charging and discharging multi-objective optimization scheduling method based on discrete action set

Abstract

The invention discloses an electric vehicle charging and discharging multi-objective optimization scheduling method based on a discrete action set. The method comprises the following steps: obtaining prediction data of a micro-grid and transmitting the prediction data to a decision center; the decision-making center inputs the prediction data into a charging and discharging multi-objective optimization scheduling model, and coordinates a complementarity power exchange process between the micro-grid and the electric vehicle by adopting a limited discrete action set and a matched auction algorithm; and obtaining an optimal prospective charging and discharging power plan of each time slice of the electric vehicle by using a decomposition-based multi-objective evolutionary algorithm so as to optimize and coordinate charging and discharging behaviors. According to the method, charge and discharge actions are discretized, so that real-time scheduling becomes possible; an auction algorithm is introduced to coordinate power exchange between the micro-grid and the electric vehicle, so that the effects of peak load shifting and interest arbitrage of a vehicle owner are achieved; and by means of a decomposition-based multi-objective evolutionary algorithm, an optimal prospective charging and discharging power plan of each time slice of the electric vehicle in an optimization period is obtained, so that the operation efficiency and the economic benefit of the micro-grid are improved.

Description

Technical Field

[0001] The invention relates to the field of electric vehicles, and in particular to a multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets. Background Art

[0002] Electric vehicles are not only a revolutionary force in the field of transportation, but also play an increasingly important role as distributed energy in smart grids with their on-board battery systems as energy storage units. With a large number of electric vehicles connected to the power grid in an unordered manner, grid load fluctuations may be aggravated, posing a huge challenge to the safety and stability of the grid. At the same time, due to the rapid development of V2G technology, electric vehicle batteries, as mobile and distributed energy storage devices, allow two-way energy exchange between electric vehicles and the power grid, that is, electric vehicles can be charged and discharged.

[0003] Under the traditional random optimization model, the continuous random charging and discharging of batteries will lead to excessive computational and communication complexity. In addition, most of the current research on optimization scheduling generally only considers the single optimization goal of load fluctuation, ignoring the interests of car owners and thus affecting their enthusiasm for participating in scheduling. Therefore, it is urgent to explore a charging and discharging scheduling method to avoid the problem of excessive computational and communication complexity caused by continuous random charging and discharging, while enabling it to alleviate grid load fluctuations and enable car owners to arbitrage when electricity market prices fluctuate, so as to achieve optimized management and arbitrage on the microgrid side. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides an electric vehicle charging and discharging multi-objective optimization scheduling method based on a discrete action set with a simple algorithm.

[0005] The technical solution of the present invention to solve the above technical problems is: a multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets, comprising the following steps:

[0006] Step 1: Obtain the forecast data of the basic load and time-of-use electricity price without considering the role of electric vehicles in each optimization cycle of the microgrid, and transmit the forecast data to the decision center;

[0007] Step 2: The decision center inputs the prediction data into the pre-built multi-objective optimization scheduling model for charging and discharging, and uses a limited discrete action set and a matching auction algorithm to coordinate the complementary power exchange process between the microgrid and electric vehicles;

[0008] Step 3: Use a decomposition-based multi-objective evolutionary algorithm with trade-off preference characteristics to obtain the optimal forward-looking charging and discharging power plan for each time slice of the electric vehicle within the optimization period to optimize the coordinated charging and discharging behavior.

[0009] The multi-objective optimization scheduling method for electric vehicle charging and discharging based on discrete action sets is described, and the specific process of step 2 is as follows:

[0010] Step S21: constructing a multi-objective optimization scheduling model for charging and discharging;

[0011] Step S22: determining the constraint conditions of the constructed charging and discharging multi-objective optimization scheduling model;

[0012] Step S23: Propose a limited set of discrete actions;

[0013] Step S24: designing an auction algorithm;

[0014] Step S25: The decision center inputs the prediction data into the charging and discharging multi-objective optimization scheduling model, and adopts a limited discrete action set and an auction algorithm to coordinate the complementary power exchange process between the microgrid and the electric vehicle.

[0015] In the above-mentioned multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets, in step S21, the main objectives of the multi-objective optimization scheduling model for charging and discharging include minimizing grid load fluctuations and maximizing arbitrage income of electric vehicle owners. The constructed multi-objective optimization scheduling model for charging and discharging is:

[0016]

[0017] MinF1 represents minimizing grid load fluctuations, MaxF2 represents maximizing arbitrage income, T represents the optimization period, n represents the number of electric vehicles, and p represents L,t represents the basic load without electric vehicles connected at time slice t, p avr represents the average load after connecting electric vehicles during the optimization period T, p i,t It represents the charging and discharging power value of the i-th electric vehicle in time slice t. A positive number represents discharging, a negative number represents charging, and 0 represents not participating in the scheduling. t represents the electricity price in time slice t, and the variable β t Determine the electricity price for time slice t as a discharge subsidy or a charging price;

[0018] Determine p as follows avr and β t :

[0019]

[0020] p avr represents the average load after connecting electric vehicles during the optimization period T, β t Determine whether the electricity price in time slice t is a discharge subsidy or a charging price. When the sum of the charging and discharging power of all electric vehicles in time slice t is a negative number, it means that the current charging power is greater than the discharging power, that is, the electricity price at this time is the charging price, βt =-1; otherwise, when the sum is positive, it is the discharge subsidy price, i.e., arbitrage income, β t =1; when the sum is 0, the charging and discharging power of the time slice is consistent, that is, the charging cost and the subsidized electricity price offset each other, and there is no arbitrage income, β t =0.

[0021] In the above-mentioned multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets, in step S22, the constraint conditions of the multi-objective optimization scheduling model for charging and discharging are determined according to the following formula:

[0022] p i,t min ≤p i,t ≤p i,t max (5)

[0023]

[0024] Formula (5) represents the constraint of the charging and discharging power of the ith electric vehicle at time slice t, p i,t min and p i,t nax are the upper and lower limits of the charging and discharging power of the ith electric vehicle at time slice t, where p i,t min is a negative value, p i,t max is a positive value;

[0025] Formula (6) indicates that the battery of an electric vehicle is prevented from being overcharged or over-discharged, that is, the battery state of charge is constrained by an upper or lower limit, SOC i,t represents the state of charge of the battery of the ith electric vehicle at time slice t, E n is the battery capacity of the ith electric car, and SOC i They represent the upper and lower limits of the battery state of charge of an electric vehicle respectively.

[0026] In the above-mentioned multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets, in step S23, the proposal of a limited discrete action set limits the optional range of charging and discharging power, as described in the following formula:

[0027]

[0028] Equations (7) and (8) represent the set of finite discrete actions that can be selected when the battery is charged and discharged, which are mapped to different SOC i,t range, where PD i,t represents the set of discrete actions that can be selected during discharge, while PC i,tis the set of discrete actions that can be selected during charging. {c, 3c, 7c} and {d, 3d, 7d} represent the finite discrete action sets for charging and discharging, respectively. The smaller the number, the smaller the charge and discharge rate.

[0029] In the above-mentioned multi-objective optimization scheduling method for electric vehicle charging and discharging based on discrete action sets, in step S24, the auction algorithm steps are as follows:

[0030] Step S241: collecting the power demand of the microgrid for peak shaving and valley filling and the selectable charging and discharging power of the electric vehicle based on a finite discrete action set;

[0031] Step S242: assigning tasks, i.e., power requirements to buyers;

[0032] Step S243: Update the buyer's charging and discharging power, that is, bid, to obtain more optimal auction results until the termination condition is met, that is, stop the auction when the difference between the total of all bids and the demand is the smallest.

[0033] In the above-mentioned multi-objective optimization scheduling method for electric vehicle charging and discharging based on discrete action sets, in step S25, in the process of coordinating the complementary power exchange between the microgrid and the electric vehicle, the buyer in the auction scenario is the electric vehicle owner, and the seller is the microgrid. The electric vehicle completes the peak shaving and valley filling goal through charging and discharging, realizes its own arbitrage income, and ensures the stability and safety of the microgrid.

[0034] In the above-mentioned multi-objective optimization scheduling method for electric vehicle charging and discharging based on discrete action sets, in step 3, the multi-objective evolutionary algorithm based on decomposition helps decision makers make wise choices according to different weight preferences in actual engineering. The steps of the multi-objective evolutionary algorithm based on decomposition are as follows:

[0035] Step S31: Input parameters, including population size N, decision vector dimension D of the problem, initial state of charge SOC of the electric vehicle battery initial , a finite set of discrete actions Set;

[0036] Step S32: generating an initialization population, generating an initial population Pop including N individuals, each individual consisting of a D-dimensional decision vector, representing a charging and discharging power plan of the electric vehicle in each time slice within the optimization period;

[0037] Step S33: using an auction algorithm to dynamically adjust the initial population Pop to obtain a candidate plan for charging and discharging power of the electric vehicle;

[0038] Step S34: Calculation of fitness value: Calculate the adjusted Pop objective function value according to the Chebyshev method function formula, and select the ideal point from it. The ideal point is composed of the optimal value obtained on each objective;

[0039] Step S35: traverse all individuals, use crossover and mutation operators to generate candidate solutions, and then use the auction algorithm to dynamically adjust the candidate solutions according to the time-of-use electricity price and constraints to ensure their feasibility;

[0040] Step S36: Update Pop using the Chebyshev function decomposition method, and calculate the optimal solution corresponding to the weight vector under N different preferences until the loop iteration stops; finally, the optimal forward-looking charging and discharging power plan for each time slice of the electric vehicle within the optimization cycle under different preferences is obtained.

[0041] In the above-mentioned multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets, in step S36, the Chebyshev method function formula is as follows:

[0042]

[0043] g te represents the set of subproblems formed by the Chebyshev method, represents the decision vector of the j-th subproblem, w j represents the weight vector corresponding to the j-th subproblem, Represents the weight value of the j-th subproblem on the m-th objective function; represents the decision vector corresponding to the j-th sub-problem, z * represents the ideal point, represents the ideal point of the mth objective function, represents the qth objective function value corresponding to the jth subproblem, q=1,…,m, and Ω represents the decision space, that is, the search space of the algorithm.

[0044] In the above-mentioned multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets, in step S36, the calculation formula of the decision vector is as follows:

[0045]

[0046] in represents the decision vector, which is the charging and discharging power plan for optimization. represents the charging and discharging power combination of all electric vehicles in time slice h, h = 1, 2, 3, ..., T, p n,h represents the charging and discharging power of the nth electric vehicle in time slice h, and the decision vector The size of dimension D is T*n.

[0047] The beneficial effects of the present invention are as follows: first, in order to avoid the communication calculation and excessive complexity problems caused by continuous random charging and discharging, the charging and discharging actions are discretized to make real-time scheduling possible; secondly, an auction algorithm is introduced to coordinate the power exchange between the microgrid and the electric vehicle to ensure efficient allocation of resources, achieve peak shaving and valley filling, and enable the car owner to arbitrage; thirdly, with the help of a decomposition-based multi-objective evolutionary algorithm (MOEA / D), the best balance point is found between minimizing grid load fluctuations and maximizing car owners' arbitrage income, that is, the optimal forward-looking charging and discharging power plan for each time slice of the electric vehicle within the optimization period is obtained, which ultimately effectively improves the operating efficiency and economic benefits of the microgrid, and lays a solid foundation for the deep integration of smart grids and electric vehicles in the future. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is the overall flow chart of the present invention.

[0049] Figure 2 This is the discrete action set diagram proposed by the present invention.

[0050] Figure 3 This is a flow chart of the auction algorithm of the present invention.

[0051] Figure 4 It is a simulation result diagram of the present invention. DETAILED DESCRIPTION

[0052] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0053] like Figure 1 As shown, a multi-objective optimization scheduling method for electric vehicle charging and discharging based on discrete action sets includes the following steps:

[0054] Step 1: Obtain forward-looking forecast data such as basic load and time-of-use electricity price without considering the role of electric vehicles in each optimization cycle of the microgrid, and transmit the forecast data to the decision center.

[0055] Step 2: The decision center inputs the predicted data into the pre-built multi-objective optimization scheduling model for charging and discharging, and adopts a limited discrete action set and a supporting auction algorithm to coordinate the complementary power exchange process between the microgrid and electric vehicles.

[0056] The specific process of step 2 is:

[0057] Step S21: construct a multi-objective optimization scheduling model for charging and discharging.

[0058] The main objectives of the multi-objective optimization scheduling model for charging and discharging include minimizing grid load fluctuations and maximizing arbitrage income for electric vehicle owners. The constructed multi-objective optimization scheduling model for charging and discharging is:

[0059]

[0060] MinF1 represents minimizing grid load fluctuations, MaxF2 represents maximizing arbitrage income, T represents the optimization period, n represents the number of electric vehicles, and p represents L,t represents the basic load without electric vehicles connected at time slice t, p avr represents the average load after connecting electric vehicles during the optimization period T, p i,t It represents the charging and discharging power value of the i-th electric vehicle in time slice t. A positive number represents discharging, a negative number represents charging, and 0 represents not participating in the scheduling. t represents the electricity price in time slice t, and the variable β t Determine the electricity price in time slice t as a discharge subsidy or a charging price.

[0061] Determine p as follows avr and β t :

[0062]

[0063] p avr represents the average load after connecting electric vehicles during the optimization period T, β t Determine whether the electricity price in time slice t is a discharge subsidy or a charging price. When the sum of the charging and discharging power of all electric vehicles in time slice t is a negative number, it means that the current charging power is greater than the discharging power, that is, the electricity price at this time is the charging price, β t =-1; otherwise, when the sum is positive, it is the discharge subsidy price, i.e., arbitrage income, β t =1; when the sum is 0, the charging and discharging power of the time slice is consistent, that is, the charging cost and the subsidized electricity price offset each other, and there is no arbitrage income, β t =0.

[0064] Step S22: Determine the constraint conditions of the constructed charging and discharging multi-objective optimization scheduling model.

[0065] The constraints of the multi-objective optimization scheduling model for charging and discharging are determined as follows:

[0066]

[0067] Formula (5) represents the constraint of the charging and discharging power of the ith electric vehicle at time slice t, p i,t min and p i,t max are the upper and lower limits of the charging and discharging power of the ith electric vehicle at time slice t, where p i,t min is a negative value, p i,t max Is a positive value.

[0068] Formula (6) indicates that the battery of an electric vehicle is prevented from being overcharged or over-discharged, that is, the battery state of charge is constrained by an upper or lower limit, SOC i,t represents the state of charge of the battery of the ith electric vehicle at time slice t, E n is the battery capacity of the ith electric car, and SOC i They represent the upper and lower limits of the battery state of charge of an electric vehicle respectively.

[0069] Step S23: Propose a limited set of discrete actions.

[0070] The proposal of a finite discrete action set limits the optional range of charge and discharge power, as described in the following formula:

[0071]

[0072] Equations (7) and (8) represent the set of finite discrete actions that can be selected when the battery is charged and discharged, which are mapped to different SOC i,t range, where PD i,t represents the set of discrete actions that can be selected during discharge, while PC i,t is the set of discrete actions that can be selected during charging. {c, 3c, 7c} and {d, 3d, 7d} represent the finite discrete action sets for charging and discharging respectively. The smaller the number, the smaller the charging and discharging rate, similar to the slow charging and fast charging gears of electric vehicles.

[0073] Specifically, Figure 2 As shown in the figure, three different charging and discharging action rates are defined, representing fast charging and discharging (7c or 7d), normal charging and discharging (3c or 3d), and slow charging and discharging (c or d). c or d represents the minimum standardized charging and discharging power of each electric vehicle in each time slice t. Among them, the choice of charging and discharging action rate directly depends on the battery state of charge and the time-of-use electricity price. The higher the state of charge, the more options for discharge action rate and the fewer options for charging action rate, and vice versa.

[0074] Step S24: Design an auction algorithm.

[0075] The purpose of using the auction algorithm is to allocate the power demand of the microgrid (seller) to the electric vehicle (buyer) to achieve peak load shifting and valley filling, ultimately ensuring the safe and stable operation of the power grid, while optimizing the arbitrage income of electric vehicle owners to stimulate their enthusiasm. Figure 3 , the auction algorithm steps are as follows:

[0076] Step S241: collecting the power demand of the microgrid for peak shaving and valley filling and the selectable charging and discharging power of the electric vehicle based on a finite discrete action set;

[0077] Step S242: assigning tasks, i.e., power requirements to buyers;

[0078] Step S243: Update the buyer's charging and discharging power, that is, bid, to obtain more optimal auction results until the termination condition is met, that is, stop the auction when the difference between the total of all bids and the demand is the smallest.

[0079] Step S25: The decision center inputs the prediction data into the charging and discharging multi-objective optimization scheduling model, and adopts a limited discrete action set and an auction algorithm to coordinate the complementary power exchange process between the microgrid and the electric vehicle.

[0080] In the process of coordinating the complementary power exchange between microgrids and electric vehicles, the buyer in the auction scenario is the electric vehicle owner, and the seller is the microgrid. Electric vehicles achieve the goal of peak shaving and valley filling through charging and discharging, realizing their own arbitrage income while ensuring the stability and safety of the microgrid.

[0081] Step 3: Use a decomposition-based multi-objective evolutionary algorithm with trade-off preference characteristics to obtain the optimal forward-looking charging and discharging power plan for each time slice of the electric vehicle within the optimization period to optimize the coordinated charging and discharging behavior.

[0082] In step 3, the multi-objective evolutionary algorithm based on decomposition helps decision makers make wise choices based on different weight preferences in actual engineering. The steps of the multi-objective evolutionary algorithm based on decomposition are as follows:

[0083] Step S31: Input parameters, including population size N, decision vector dimension D of the problem, initial state of charge SOC of the electric vehicle battery initial , a finite set of discrete actions Set.

[0084] Step S32: Generate an initialization population, generate an initial population Pop including N individuals, each individual is composed of a D-dimensional decision vector, representing the charging and discharging power plan of the electric vehicle in each time slice within the optimization period.

[0085] Step S33: using the auction algorithm to dynamically adjust the initial population Pop to obtain candidate charging and discharging power plans for electric vehicles.

[0086] Step S34: Calculation of fitness value: Calculate the adjusted Pop objective function value according to the Chebyshev method function formula, and select the ideal point from it. The ideal point is composed of the optimal value obtained on each objective. For example, in the dual-objective minimization problem, the objective function values ​​obtained by the two individuals are divided into (2,5) and (3,1). Then the ideal point z at this time is * =(2,1).

[0087] Step S35: traverse all individuals, use crossover and mutation operators to generate candidate solutions, and then use the auction algorithm to dynamically adjust the candidate solutions according to the time-of-use electricity price and constraints to ensure their feasibility.

[0088] Step S36: Update Pop using the Chebyshev function decomposition method, and calculate the optimal solution corresponding to the weight vector under N different preferences until the loop iteration stops; finally, the optimal forward-looking charging and discharging power plan for each time slice of the electric vehicle within the optimization cycle under different preferences is obtained.

[0089] In step S36, the Chebyshev method function formula is as follows:

[0090]

[0091] g te represents the set of subproblems formed by the Chebyshev method, represents the decision vector of the j-th subproblem, w j represents the weight vector corresponding to the j-th subproblem, Represents the weight value of the j-th subproblem on the m-th objective function; represents the decision vector corresponding to the j-th sub-problem, z * represents the ideal point, represents the ideal point of the mth objective function, represents the qth objective function value corresponding to the jth subproblem, q=1,…,m, and Ω represents the decision space, that is, the search space of the algorithm.

[0092] The calculation formula of the decision vector is as follows:

[0093]

[0094] in represents the decision vector, which is the charging and discharging power plan for optimization. represents the charging and discharging power combination of all electric vehicles in time slice h, h = 1, 2, 3, ..., T, p n,h represents the charging and discharging power of the nth electric vehicle in time slice h, and the decision vector The size of dimension D is T*n.

[0095] like Figure 4 As shown in the figure, through the multi-objective optimization scheduling method of electric vehicle charging and discharging based on discrete action sets, it can be obtained that: the bar charts above and below the horizontal axis represent the total charging power and total discharging power of all vehicles participating in the charging and discharging scheduling in each time slice, and the dotted line represents the electricity price. The simulation results show that when it is during the peak period of electricity consumption, that is, Figure 4In the 10:00-13:00 and 18:00-21:00 time slices, when the electricity price is high, electric vehicles will discharge more frequently to obtain arbitrage income, and at the same time balance the grid load fluctuations. When the electricity price is low, electric vehicles will charge more, which also reflects the "buy low and sell high" strategy in the auction algorithm. However, electric vehicles participating in charging and discharging scheduling must not only consider arbitrage income, but also grid load fluctuations. The simulation results show that in some time slices (for example, 1, 2, 3, 4, 5, etc.), some electric vehicles are charging, while others are discharging, which is in line with the actual situation.

Claims

1. A multi-objective optimization scheduling method for electric vehicle charging and discharging based on discrete action sets, It is characterized in that The following steps are involved: Step 1: Obtain the forecast data of the basic load and time-of-use electricity price without considering the role of electric vehicles in each optimization cycle of the microgrid, and transmit the forecast data to the decision center; Step 2: The decision center inputs the prediction data into the pre-built multi-objective optimization scheduling model for charging and discharging, and uses a limited discrete action set and a matching auction algorithm to coordinate the complementary power exchange process between the microgrid and electric vehicles; Step 3: Use a decomposition-based multi-objective evolutionary algorithm with trade-off preference characteristics to obtain the optimal forward-looking charging and discharging power plan for each time slice of the electric vehicle within the optimization period to optimize the coordinated charging and discharging behavior.

2. The multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets according to claim 1, It is characterized in that The specific process of step 2 is: Step S21: constructing a multi-objective optimization scheduling model for charging and discharging; Step S22: determining the constraint conditions of the constructed charging and discharging multi-objective optimization scheduling model; Step S23: Propose a limited set of discrete actions; Step S24: designing an auction algorithm; Step S25: The decision center inputs the prediction data into the charging and discharging multi-objective optimization scheduling model, and adopts a limited discrete action set and an auction algorithm to coordinate the complementary power exchange process between the microgrid and the electric vehicle.

3. The multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets according to claim 2, It is characterized in that In step S21, the main objectives of the charging and discharging multi-objective optimization scheduling model include minimizing grid load fluctuations and maximizing arbitrage income of electric vehicle owners. The constructed charging and discharging multi-objective optimization scheduling model is: MinF 1 Indicates minimizing grid load fluctuation, MaxF 2 represents the maximum arbitrage income, T represents the optimization period, n represents the number of electric vehicles, p L，t represents the basic load without electric vehicles connected at time slice t, pav r represents the average load after connecting electric vehicles during the optimization period T, p i，t It represents the charging and discharging power value of the i-th electric vehicle in time slice t. A positive number represents discharging, a negative number represents charging, and 0 represents not participating in the scheduling. t represents the electricity price in time slice t, and the variable β t Determine the electricity price for time slice t as a discharge subsidy or a charging price; Determine p as follows avr and β t : p avr represents the average load after connecting electric vehicles during the optimization period T, β t Determine whether the electricity price in time slice t is a discharge subsidy or a charging price. When the sum of the charging and discharging power of all electric vehicles in time slice t is a negative number, it means that the current charging power is greater than the discharging power, that is, the electricity price at this time is the charging price, β t =-1; otherwise, when the sum is positive, it is the discharge subsidy price, i.e., arbitrage income, β t =1; when the sum is 0, the charging and discharging power of the time slice is consistent, that is, the charging cost and the subsidized electricity price offset each other, and there is no arbitrage income, β t =0.

4. The multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets according to claim 2, It is characterized in that In step S22, the constraint conditions of the multi-objective optimization scheduling model for charging and discharging are determined as follows: p i，t min ≤p i,t ≤p i，t max (5) Formula (5) represents the constraint of the charging and discharging power of the ith electric vehicle at time slice t, p i，t min and p i，t max are the upper and lower limits of the charging and discharging power of the ith electric vehicle at time slice t, where p i，t min is a negative value, p i，t max is a positive value; Formula (6) indicates that the battery of an electric vehicle is prevented from being overcharged or over-discharged, that is, the battery state of charge is constrained by an upper or lower limit, SOC i，t represents the state of charge of the battery of the ith electric vehicle at time slice t, E n is the battery capacity of the ith electric car, and SOC i They represent the upper and lower limits of the battery state of charge of an electric vehicle respectively.

5. The multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets according to claim 2, It is characterized in that In step S23, the proposal of a limited discrete action set limits the optional range of charging and discharging power, as described in the following formula: Equations (7) and (8) represent the set of finite discrete actions that can be selected when the battery is charged and discharged, which are mapped to different SOC i，t range, where PD i,t represents the set of discrete actions that can be selected during discharge, while PC i，t It is the set of discrete actions that can be selected during charging. {c, 3c, 7c} and {d, 3d, 7d} represent the finite discrete action sets for charging and discharging respectively. The smaller the number, the smaller the charge and discharge rate.

6. The multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets according to claim 2, It is characterized in that In step S24, the auction algorithm steps are as follows: Step S241: collecting the power demand of the microgrid for peak shaving and valley filling and the selectable charging and discharging power of the electric vehicle based on a finite discrete action set; Step S242: assigning tasks, i.e., power requirements to buyers; Step S243: Update the buyer's charging and discharging power, that is, bid, to obtain more optimal auction results until the termination condition is met, that is, stop the auction when the difference between the total of all bids and the demand is the smallest.

7. The multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets according to claim 2, It is characterized in that In step S25, in the process of coordinating the complementary power exchange between the microgrid and the electric vehicle, the buyer in the auction scenario is the electric vehicle owner, and the seller is the microgrid. The electric vehicle completes the peak load reduction and valley load filling goal through charging and discharging, realizes its own arbitrage income, and ensures the stability and safety of the microgrid.

8. The multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets according to claim 1, It is characterized in that In step 3, the multi-objective evolutionary algorithm based on decomposition helps decision makers make wise choices based on different weight preferences in actual engineering. The steps of the multi-objective evolutionary algorithm based on decomposition are as follows: Step S31: Input parameters, including population size N, decision vector dimension D of the problem, initial state of charge SOC of the electric vehicle battery initial , a finite set of discrete actions Set; Step S32: generating an initialization population, generating an initial population Pop including N individuals, each individual consisting of a D-dimensional decision vector, representing a charging and discharging power plan of the electric vehicle in each time slice within the optimization period; Step S33: using an auction algorithm to dynamically adjust the initial population Pop to obtain a candidate plan for charging and discharging power of the electric vehicle; Step S34: Calculation of fitness value: Calculate the adjusted Pop objective function value according to the Chebyshev method function formula, and select the ideal point from it. The ideal point is composed of the optimal value obtained on each objective; Step S35: traverse all individuals, use crossover and mutation operators to generate candidate solutions, and then use the auction algorithm to dynamically adjust the candidate solutions according to the time-of-use electricity price and constraints to ensure their feasibility; Step S36: Update Pop using the Chebyshev function decomposition method, and calculate the optimal solution corresponding to the weight vector under N different preferences until the loop iteration stops; finally, the optimal forward-looking charging and discharging power plan for each time slice of the electric vehicle within the optimization cycle under different preferences is obtained.

9. The multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets according to claim 8, It is characterized in that In step S36, the Chebyshev method function formula is as follows: g te represents the set of subproblems formed by the Chebyshev method, represents the decision vector of the j-th subproblem, w j represents the weight vector corresponding to the j-th subproblem, Represents the weight value of the j-th subproblem on the m-th objective function; represents the decision vector corresponding to the j-th sub-problem, z * represents the ideal point, represents the ideal point of the mth objective function, represents the qth objective function value corresponding to the jth subproblem, q = 1, ..., m, and Ω represents the decision space, that is, the search space of the algorithm.

10. The multi-objective optimization scheduling method for charging and discharging electric vehicles based on discrete action sets according to claim 9, It is characterized in that In step S36, the calculation formula of the decision vector is as follows: in represents the decision vector, which is the charging and discharging power plan for optimization. represents the charging and discharging power combination of all electric vehicles in time slice h, h = 1, 2, 3, ..., T, p n，h represents the charging and discharging power of the nth electric vehicle in time slice h, and the decision vector The size of dimension D is T*n.

Images