An electric vehicle charging and discharging scheduling method based on a simulation strategy improvement algorithm

Abstract

The application discloses a kind of based on simulation strategy improvement algorithm electric vehicle charging and discharging scheduling method, including the following steps: establishing the multi-stage charging and discharging scheduling control Markov decision model considering the random charging demand of electric vehicle, determine model constraint condition and optimization target;Based on Monte Carlo simulation simulation generates the future charging demand sample path of electric vehicle;Combining the problem structure features, the feasible charging and discharging behavior set of electric vehicle is reduced screening;The Q factor of each feasible behavior is calculated using the simulation-based strategy improvement algorithm, and the basic strategy is updated to obtain the improved strategy.The charging and discharging scheduling strategy obtained by the application considers the influence of multi-period charging and discharging control, can adapt to the change of future uncertain charging demand, and can improve the solution speed and efficiency of electric vehicle charging and discharging scheduling control.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electric vehicle charging and discharging scheduling control, in particular to an electric vehicle charging scheduling method based on simulation strategy improvement algorithm. BACKGROUND

[0002] With the continuous aggravation of environmental pollution, electric vehicles have received widespread attention in recent years. Electric vehicles are driven by on-board power sources, have the advantages of low pollution and fuel cost savings. The large-scale popularity of electric vehicles can reduce energy consumption and reduce tail gas emissions, thereby achieving the purpose of energy saving and emission reduction. It can be predicted that the future electric vehicle market will maintain a strong upward trend, and electric vehicles will gradually achieve large-scale popularity. Large-scale popularity of electric vehicles will be an important means to reduce consumption of fossil energy and alleviate environmental pollution.

[0003] Although the large-scale popularity of electric vehicles can reduce environmental pollution, the large number of newly emerging electric vehicle charging and discharging demands will have a great impact on the power grid. How to reasonably and efficiently schedule and control electric vehicle charging and discharging demands will be a key factor affecting the popularity of electric vehicles. Electric vehicle charging and discharging scheduling control mainly faces the following difficulties. First, electric vehicle charging demand has randomness. The start time, end time and required charging energy of electric vehicle charging have certain uncertainty. Second, charging and discharging control needs to consider multi-period influence. Due to the time coupling of charging demand, current charging and discharging behavior will affect future charging demand, so multi-stage charging and discharging control is needed. Finally, the charging and discharging scheduling strategy space grows exponentially with the number of electric vehicles. A large number of electric vehicles will bring large state space and large action space problems, which are difficult to solve efficiently. SUMMARY

[0004] In view of the above difficulties in electric vehicle charging and discharging scheduling, the present application proposes an electric vehicle charging and discharging scheduling method based on simulation strategy improvement. Through Monte Carlo simulation estimation and online strategy improvement, combined with the structural characteristics of the problem, the problem of uncertain demand, multi-period decision and large state and action space is overcome.

[0005] The technical scheme for achieving the object of the present application is: an electric vehicle charging and discharging scheduling method based on simulation strategy improvement algorithm, comprising the following steps:

[0006] Step 1, obtain the total number of electric vehicles, the start time of the parking charging demand of the electric vehicles obeys a probability distribution, the parking time obeys a probability distribution, and the required charging energy obeys a probability distribution;

[0007] Step 2, establish a multi-stage charging and discharging scheduling control Markov decision model considering the random charging demand of electric vehicles, and determine the constraint conditions of charging and discharging scheduling;

[0008] Step 3, a Monte Carlo simulation method is used to simulate the future charging demand of each electric vehicle to generate a plurality of sample paths;

[0009] Step 4, according to the state of the electric vehicle to be charged at the current time, the flip charging and discharging behavior is deleted, and the set of charging and discharging behaviors to be evaluated is determined;

[0010] Step 5, the Q factor of each charging and discharging behavior is calculated based on the basic strategy; the behavior with the maximum Q factor value is selected as the current optimal charging and discharging control behavior, and the performance of the current basic strategy is improved.

[0011] An electronic device includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor implements the electric vehicle charging and discharging scheduling method based on the simulation strategy improvement algorithm described above when executing the program.

[0012] A computer readable storage medium has a computer program stored thereon, and the program is executed by a processor to implement the electric vehicle charging and discharging scheduling method based on the simulation strategy improvement algorithm described above.

[0013] The above technical scheme is adopted, and the following technical effects can be achieved:

[0014] The Markov decision model of the electric vehicle charging and discharging scheduling control established can effectively represent the future uncertain charging demand of the electric vehicle, and the influence of multi-time charging and discharging decision on the charging and discharging cost is considered, so that the adaptability of the charging and discharging scheduling strategy to uncertain charging demand can be improved. Based on the above Markov decision model, a simulation-based strategy improvement algorithm is proposed, which can effectively avoid the occurrence of flip charging situation, reduce user dissatisfaction, and at the same time, the algorithm does not need to perform iterative optimization calculation, compared with the traditional strategy iteration and value iteration method, the algorithm can effectively solve the difficulty of large state space and large action space, and improve the solving speed and efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 is the principle block diagram of the electric vehicle charging and discharging scheduling method of the application.

[0016] Figure 2 is a schematic diagram of the charging behavior of the electric vehicle after the charging and discharging scheduling. DETAILED DESCRIPTION

[0017] The technical scheme of the application will be further described below in combination with the drawings and examples.

[0018] As Figure 1As shown, this invention mainly includes: an initialization module, a multi-stage stochastic charge-discharge scheduling model construction module, a constraint condition construction module, an optimization objective function construction module, a sample path simulation module, a behavior set screening module, an optimal Q-factor calculation module, and an optimal charge-discharge scheduling strategy extraction and output module. The specific implementation steps are illustrated in the flowchart below:

[0019] Step 1: Initialization. The operator inputs the number N of electric vehicles to be scheduled for charging and discharging. The start time of an electric vehicle's parking and charging demand follows a probability distribution P. b The parking duration follows a probability distribution P. d The required charging energy follows a probability distribution P. e ;

[0020] Step 2: Constructing a multi-stage stochastic charge / discharge scheduling model. Define the model state as follows: in, This represents the remaining dwell time of the i-th electric vehicle at time t. Let represent the i-th electric vehicle at time t. Define the model action as... in, Let represent the charging and discharging behavior of the i-th electric vehicle at time t. This indicates that the i-th electric vehicle is being charged at this time. This indicates that a discharge is being performed. This indicates that no charging or discharging operation will be performed. (Based on the electric vehicle's status.) The dynamic equations of the system in the model can be expressed as:

[0021]

[0022]

[0023] Where ΔT represents the time interval between two adjacent decision moments. According to the probability distribution P d The simulated parking duration, where P represents the charging and discharging power, and ψ... c Indicates charging efficiency, ψ d Indicates discharge efficiency. According to the probability distribution P e The required charging energy generated by the simulation.

[0024] Step 3: Constraint Construction. Due to the need to meet the charging requirements of electric vehicles, and considering the physical characteristics of electric vehicle batteries, the following constraint equations must be satisfied during the charging and discharging process of the electric vehicle:

[0025]

[0026]

[0027]

[0028] wherein formula (1) indicates that the remaining charging energy of the electric vehicle cannot be greater than the battery capacity E of the electric vehicle cap , formula (2) indicates that the remaining charging energy of the electric vehicle cannot be greater than the maximum charging energy that can be provided within the remaining parking period. Formula (3) indicates the feasible value range of the charging and discharging behavior of the electric vehicle under different conditions.

[0029] Step 4: optimization objective function construction. Considering that minimizing the charging cost of the electric vehicle is taken as the optimization objective of the charging and discharging scheduling of the electric vehicle, there is the following equation:

[0030]

[0031] wherein J represents the optimization objective value of the problem, π represents the charging and discharging scheduling strategy to be optimized, T represents the number of decision cycles, ω t represents the charging and discharging price of the electric vehicle, and E represents the expected operation.

[0032] Step 5: sample path simulation. For each electric vehicle, firstly, the starting time t b of the future new parking charging demand event is sampled based on the probability distribution P next of the starting time of the parking charging demand of the electric vehicle. Secondly, the duration d next of the future new parking charging demand event is sampled based on the sampling generated starting time t d and the parking duration probability distribution P next . Finally, the energy demand e e of the future new parking charging demand event is sampled based on the required charging energy probability distribution P next .

[0033] The triple [t next , d next , e next ] is recorded as a sample path of the future charging demand of the i-th electric vehicle The above process is repeated until a total of M sample paths are generated for each electric vehicle.

[0034] Step 6: behavior set screening. Firstly, the feasible value range of the decision behavior of each electric vehicle is determined according to formula (5) at the current decision time t. Secondly, the decision action of each electric vehicle at t-1 time is read, and if the t-1 time , then can only take the value 0 or 1. If ​but It can only take the value 0 or -1. This reduces the frequency of electric vehicles tipping over to recharge, thus lowering user dissatisfaction. All feasible decision behaviors of electric vehicles are combined to form the set of charging and discharging behaviors to be evaluated.

[0035] Step 7: Calculate the optimal Q factor

[0036] The existing electric vehicle charging and discharging scheduling strategy is selected as the basic strategy π. b For example, the following greedy charging strategy can be selected as the basic strategy.

[0037] if

[0038] For each behavior in the set of charge and discharge behaviors to be evaluated Calculate the Q-factor for each behavior, as shown in the equation below.

[0039]

[0040] in, This indicates that based on the basic strategy π b It determines the charging and discharging behavior of every electric vehicle at every moment in the future. This indicates that the future charging and discharging cost is calculated based on the generated m-th sample path.

[0041] Select behavior A with the largest Q factor t As the optimization behavior at the current time t, the basic policy is updated at time t, and the state S is... t The optimal behavior under these conditions is shown by the following equation:

[0042]

[0043] in, This represents the set of charging and discharging behaviors to be evaluated.

[0044] Step 8:

[0045] Update the base strategy π according to formula (9) b In state S t The following actions are used to obtain an improved strategy for electric vehicle charging and discharging scheduling, and the output is proposed.

[0046] The experimental results of this invention are as follows: Figure 2 As shown in the figure. The hardware environment is: CPU i7-11700K@3.60GHz, memory 16G, and the charging and discharging process of 10 electric vehicles is optimized and scheduled. Figure 2 The results of electric vehicle charging and discharging scheduling are presented. Figure 2The lower half of the figure shows the parking state of the electric vehicle, and the vertical coordinate of 1 indicates that the electric vehicle is in the parking charging state, and the vertical coordinate of 0 indicates that the electric vehicle is in the driving state. Figure 2 The upper half of the figure shows the charging and discharging scheduling behavior of the electric vehicle, and the vertical coordinate of 1 indicates that the electric vehicle is charging at this moment, the vertical coordinate of -1 indicates that the electric vehicle is discharging at this moment, and the vertical coordinate of 0 indicates that the electric vehicle does not perform any charging and discharging operation at this moment. Figure 2 It can be seen that after scheduling optimization, the charging behavior of the electric vehicle mainly occurs at the low electricity price time in the morning, and the discharging behavior mainly occurs at the high electricity price time at night, so as to reduce the charging and discharging cost of the electric vehicle.

[0047] The multi-stage random charging and discharging scheduling model of the application can effectively represent the future uncertain charging demand and solve the time coupling difficulty of multi-stage charging and discharging joint scheduling. When solving the model, through the proposed sample path simulation and strategy improvement algorithm based on Monte Carlo simulation, combined with the problem characteristics to filter the feasible charging and discharging behavior set, the large state space and large action space difficulties in the model solving process can be solved, and the problem solving efficiency is improved.

[0048] The application provides an electric vehicle charging and discharging scheduling method based on a simulation strategy improvement algorithm, and there are many methods and ways to realize the technical scheme, and the above description is only the preferred embodiment of the application, and it should be pointed out that for ordinary skilled persons in the technical field, some improvements and decorations can be made without departing from the principle of the application, and these improvements and decorations should be regarded as the protection scope of the application. The components not explicitly described in the embodiment can be realized by using the existing technology.

Claims

1. A method for electric vehicle charging and discharging scheduling based on a simulation strategy improvement algorithm, characterized in that, The method comprises the following steps: Step 1, obtaining the total number of electric vehicles, the starting time of the electric vehicle parking charging demand obeying a probability distribution, the parking time obeying a probability distribution, and the required charging energy obeying a probability distribution; Step 2, establishing a Markov decision model of multi-stage charging and discharging scheduling control considering the random charging demand of electric vehicles, and determining the constraint conditions of charging and discharging scheduling; Step 3, using the Monte Carlo simulation method to simulate and generate multiple sample paths of future charging demand for each electric vehicle; Step 4, according to the state of the electric vehicle to be charged at the current time, deleting the flip charging and discharging behavior, and determining the set of charging and discharging behaviors to be evaluated; Step 5, calculating the Q factor of each charging and discharging behavior based on the basic strategy; selecting the behavior with the minimum Q factor value as the current optimal charging and discharging control behavior, and improving the performance of the current basic strategy, the specific steps are as follows: Step 5-1, select an existing electric vehicle charging and discharging scheduling strategy as the base strategy π b ; Step 5-2, for each behavior in the set of behaviors to be evaluated represents the charging and discharging behavior of the i-th electric vehicle at time t, represents charging the i-th electric vehicle at this time, represents discharging, represents not performing any charging and discharging operation, N is the total number of electric vehicles; calculate the Q factor of each behavior as shown in the following equation: wherein P represents the charging and discharging power, T represents the number of decision cycles, ω t represents the electric vehicle charging and discharging price, is a sample path of the future charging demand of the i-th electric vehicle, and M sample paths are generated for each electric vehicle; represents the future charging and discharging behavior of each electric vehicle at each time according to the basic strategy π b represents the future charging and discharging behavior of each electric vehicle at each time according to the basic strategy π represents the future charging and discharging cost calculated according to the generated m-th sample path; Step 5 - 3, select the action A with the smallest Q-factor t As the optimized action at the current time t, the base policy is updated at time t, the optimal action under the state S t is shown in the following equation: wherein, represents the set of charging and discharging behaviors to be evaluated; the improved strategy of the electric vehicle charging and discharging scheduling can be obtained through formula (8).

2. The electric vehicle charging and discharging scheduling method based on the simulation strategy improvement algorithm according to claim 1, characterized in that, Step 2 comprises: Step 2-1, the state definition of the Markov decision model is defined as wherein, represents the remaining stay time of the i-th electric vehicle at time t, represents the remaining charging energy of the i-th electric vehicle at time t; the action is defined as Step 2-2, the following system dynamic equation is established: where ΔT denotes the time interval between two consecutive decision instants, denotes the probability distribution P d simulated generated parking duration, ψ c denotes the charging efficiency, ψ d denotes the discharging efficiency, denotes the probability distribution P e simulated generated required charging energy; Step 2-3, the following constraint equation is established: wherein formula (3) represents that the remaining charging energy required by the electric vehicle cannot be greater than the battery capacity E of the electric vehicle cap , formula (4) represents that the remaining charging energy required by the electric vehicle cannot be greater than the maximum charging energy that can be provided within the remaining parking period, and formula (5) represents the feasible value range of the charging and discharging behavior of the electric vehicle under different conditions. Step 2-4, the following optimization objective equation is established: Wherein, J represents the optimization objective value of the problem, π represents the electric vehicle charging and discharging scheduling strategy to be optimized, and E represents the expectation operation.

3. The electric vehicle charging and discharging scheduling method based on the simulation strategy improvement algorithm according to claim 2, characterized in that, Step 3 comprises: Step 3-1, for the ith electric vehicle, based on the probability distribution P of the starting time of the parking charging demand of the electric vehicle b , sample the starting time t of the future occurrence of a new parking charging demand event next ; Step 3-2, for the ith electric vehicle, based on the sampled generation start time t next and parking duration probability distribution P d , sample the duration d next of the future new parking charging demand event. Step 3-3, for the i-th electric vehicle, based on the required charging energy probability distribution P e , sample the energy demand e next for a future occurrence of a new parking charging demand event. Step 3-4, the triple [t next ,d next ,e next ] is denoted as a sample path of the future charging demand of the i-th electric vehicle Steps 3-1, 3-2 and 3-3 are repeated until M sample paths are generated for each electric vehicle in total.

4. The electric vehicle charging and discharging scheduling method based on the simulation strategy improvement algorithm according to claim 3, characterized in that, Step 4 comprises: Step 4-1, for the current decision time t, determine the feasible value range of each electric vehicle decision behavior according to formula (5) ; Step 4-2, read each electric vehicle's decision action at time t-1, if t-1 time then can only take the value 0 or 1; if then can only take the value 0 or -1; all electric vehicle's feasible decision action combinations together form the set of charging and discharging behaviors to be evaluated.

5. The electric vehicle charging and discharging scheduling method based on the simulation strategy improvement algorithm according to claim 1, characterized in that, In step 5-1, the following greedy charging strategy is selected as the basic strategy 6. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to realize the electric vehicle charging and discharging scheduling method based on the simulation strategy improvement algorithm according to any one of claims 1-5.

7. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to realize the electric vehicle charging and discharging scheduling method based on the simulation strategy improvement algorithm according to any one of claims 1-5.

Images