Distributed intelligent power distribution optimization method, device, storage medium and terminal equipment

Abstract

The present application provides a distributed intelligent power distribution optimization method, device, storage medium and terminal equipment, which relates to the field of intelligent power distribution optimization technology. The method includes: constructing a mean field distribution model that characterizes the probability distribution of different business states of each smart grid entity in a distributed smart grid; constructing a revenue function of the smart grid entity based on the mean field distribution model, performing dynamic programming transformation on the revenue function, and obtaining a value function, wherein the power distribution strategy includes the business actions of the smart grid entity under the current business state and time; using the value function as the control equation, solving the control equation based on a preset multi-objective dynamic optimization algorithm, and solving the preset mean field evolution equation based on the finite difference method, so as to obtain the optimal power distribution strategy of each smart grid entity when both the control equation and the mean field evolution equation reach the Nash equilibrium point, and performing power distribution optimization based on the optimal power distribution strategy. The present application can significantly improve the operating efficiency of the power grid.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent power distribution optimization, and specifically to a distributed intelligent power distribution optimization method, a distributed intelligent power distribution optimization device, a machine-readable storage medium and a terminal device. Background Art

[0002] The development of the power system field is accompanied by the demand for sustainable development. The importance of new power systems is growing. The future power production and supply system is called smart grid, which will bring significant benefits to energy production and dispatch. However, the current smart grid has the problems of high complexity, high dynamics and high heterogeneity of the distribution system. At the same time, most new energy sources are restricted by geographical location, climate change and weather factors, which makes the supply balance of the power system more stressful than that of the traditional power grid. The optimization of the distribution network faces huge challenges. At present, based on the comprehensive consideration of the computing power of smart grid entities, the communication transmission efficiency of distribution tasks and the communication transmission price, classical game theory is usually used to dynamically model the utility function of distributed network entities in the highly dynamic heterogeneous network environment in smart grids. However, as the number of participants increases, the classical game theory model is difficult to work due to the complexity of calculation and cannot accurately predict the optimal distribution strategy of smart grids. Summary of the invention

[0003] The purpose of the embodiments of the present application is to provide a distributed intelligent power distribution optimization method, device, storage medium and terminal device to solve the above-mentioned problems.

[0004] In order to achieve the above objectives, the present application provides a first aspect of a distributed intelligent power distribution optimization method, comprising:

[0005] Construct a mean field distribution model to characterize the probability distribution of different service states of each smart grid entity in a distributed smart grid;

[0006] Constructing a revenue function of a smart grid entity based on the mean field distribution model, performing dynamic programming transformation on the revenue function, and obtaining a value function that characterizes a mapping relationship between revenue and power distribution strategy of the smart grid entity under different business states and times, wherein the power distribution strategy includes business actions of the smart grid entity under the current business state and time;

[0007] The value function is used as the control equation, the control equation is solved based on a preset multi-objective dynamic optimization algorithm, and the preset mean field evolution equation is solved based on the finite difference method, so as to obtain the optimal power distribution strategy of each smart grid entity when the control equation and the mean field evolution equation both reach the Nash equilibrium point, and the power distribution of each smart grid entity is optimized based on the optimal power distribution strategy.

[0008] Optionally, constructing a revenue function of a smart grid entity based on the average field distribution model comprises:

[0009] According to the preset business revenue factor, average field impact factor, transmission cost factor, computing cost factor, communication transmission resources and computing resources of smart grid entities, a cost-benefit function corresponding to the business status, business action and average field distribution of smart grid entities is constructed;

[0010] According to the preset penalty factor, the total amount of business tasks corresponding to the current power distribution task, the current business status of each smart grid entity and the number of smart grid entities, a penalty function representing the terminal penalty of each smart grid entity under different business task allocation states is constructed;

[0011] A benefit function of a smart grid entity is constructed based on the cost benefit function and the penalty function.

[0012] Optionally, performing dynamic programming transformation on the revenue function to obtain a value function that characterizes the mapping relationship between the revenue and the power distribution strategy of the smart grid entity under different business states and times includes:

[0013] Based on the Bellman optimality principle, the revenue function is converted into a value function of a smart grid entity;

[0014] After obtaining the value function representing the mapping relationship between the revenue and the power distribution strategy of the smart grid entity under different business states and times, the method further includes:

[0015] Converting the value function based on a preset time variation, and performing Taylor expansion on the converted value function to obtain a corresponding objective function;

[0016] The time variation is made to tend to a preset value, so as to generate a HJB equation representing the power distribution strategy and business status of the smart grid entity based on the objective function.

[0017] Optionally, the value function is used as a control equation, including:

[0018] Taking the HJB equation as the control equation;

[0019] The multi-objective dynamic optimization algorithm comprises:

[0020] S1. Initialize algorithm parameters, where the algorithm parameters include at least the number of particles in the population, the inertia weight of the particles, the initial velocity of the particles, and the initial position of the particles, wherein each particle in the population represents a feasible solution to the current power distribution task;

[0021] S2. Calculate the fitness value of each particle based on a preset fitness function;

[0022] S3. Determine the individual extreme value of each particle, the population extreme value and the average fitness value of the population according to the fitness value of each particle, wherein the individual extreme value is the optimal position searched by a single particle up to the present moment, and the population extreme value is the optimal position searched by the entire particle group up to the present moment;

[0023] S4, updating the inertia weight of each particle based on the average fitness value of the population, the maximum value of the inertia weight of the particle, and the minimum value of the inertia weight of the particle;

[0024] S5, update the current speed and current position of each particle according to the inertia weight, individual extreme value and population extreme value of each particle, return to step S2, if the population extreme value is not updated after N consecutive iterations, update the current speed and current position of each particle according to the individual extreme value and population extreme value;

[0025] S6. Determine whether the termination condition is met. If so, output the current optimal individual extreme value and the population extreme value, whichever has the larger fitness value, as the current optimal solution. Otherwise, return to step S2.

[0026] Optionally, the inertia weight of each particle is updated based on the average fitness value of the population, the maximum value of the inertia weight of the particle, and the minimum value of the inertia weight of the particle, including:

[0027] Update the inertia weight of each particle using the following formula:

[0028]

[0029] in, represents the inertia weight of the particle, Indicates the minimum value of the particle's inertia weight, Indicates the maximum value of the particle's inertia weight, represents the average fitness value of the population, g i ( t )express t Particle at time i Fitness value, min g i ( t ) represents the minimum fitness value of all particles.

[0030] Optionally, the current speed and current position of each particle are updated according to the individual extreme value and the population extreme value, including:

[0031] Update the current speed of each particle using the following formula:

[0032]

[0033] Update the current position of each particle using the following formula:

[0034]

[0035] in, Filial v represents the current velocity of the particle, Filial x represents the current position of the particle, c represents the learning factor, Pool v (seed i ) represents the speed of individual extreme values, Pool v (seed j ) represents the speed of the population extreme value, Pool x (seed i ) represents the location of individual extreme values, Pool x (seed j ) Indicates the location of the population extreme value, norm Indicates the modulus length.

[0036] Optionally, the mean field evolution equation includes:

[0037]

[0038] in, Represents the smart grid entity at time t and business status s The average field distribution under m ( t , s )express m ( t , s ) about the rate of change of energy state E, express t Energy usage status at all times, m ( t , s )express m ( t , s ) About queue status q The rate of change, express t The queue status at the moment.

[0039] Optionally, solving a preset mean field evolution equation based on a finite difference method includes:

[0040] Based on the finite difference method, the preset mean field evolution equation is solved and the following relationship is obtained:

[0041]

[0042] in, , , , Indicates time ,energy level and status The value of the mean field at Indicates time ,energy level and status The power at Indicates time ,energy level and status The value of the queue status at .

[0043] In a second aspect of the present application, a distributed intelligent power distribution optimization device is provided, comprising:

[0044] A model building module is configured to build a mean field distribution model representing the probability distribution of different business states of each smart grid entity in the distributed smart grid;

[0045] A revenue function construction module is configured to construct a revenue function of a smart grid entity based on the mean field distribution model, perform dynamic programming transformation on the revenue function, and obtain a value function that characterizes a mapping relationship between revenue and power distribution strategy of the smart grid entity under different business states and times, wherein the power distribution strategy includes a business action of the smart grid entity under a current business state and time;

[0046] The power distribution optimization module is configured to use the value function as the control equation, solve the control equation based on a preset multi-objective dynamic optimization algorithm, and solve the preset mean field evolution equation based on a finite difference method, so as to obtain the optimal power distribution strategy of each smart grid entity when the control equation and the mean field evolution equation both reach a Nash equilibrium point, and optimize the power distribution of each smart grid entity based on the optimal power distribution strategy.

[0047] Optionally, constructing a revenue function of a smart grid entity based on the average field distribution model comprises:

[0048] According to the preset business revenue factor, average field impact factor, transmission cost factor, computing cost factor, communication transmission resources and computing resources of smart grid entities, a cost-benefit function corresponding to the business status, business action and average field distribution of smart grid entities is constructed;

[0049] According to the preset penalty factor, the total amount of business tasks corresponding to the current power distribution task, the current business status of each smart grid entity and the number of smart grid entities, a penalty function representing the terminal penalty of each smart grid entity under different business task allocation states is constructed;

[0050] A benefit function of a smart grid entity is constructed based on the cost benefit function and the penalty function.

[0051] Optionally, performing dynamic programming transformation on the revenue function to obtain a value function that characterizes the mapping relationship between the revenue and the power distribution strategy of the smart grid entity under different business states and times includes:

[0052] Based on the Bellman optimality principle, the revenue function is converted into a value function of a smart grid entity;

[0053] After obtaining the value function representing the mapping relationship between the revenue and the power distribution strategy of the smart grid entity under different business states and times, the revenue function construction module is further configured as follows:

[0054] Converting the value function based on a preset time variation, and performing Taylor expansion on the converted value function to obtain a corresponding objective function;

[0055] The time variation is made to tend to a preset value, so as to generate a HJB equation representing the power distribution strategy and business status of the smart grid entity based on the objective function.

[0056] Optionally, the value function is used as a control equation, including:

[0057] Taking the HJB equation as the control equation;

[0058] The multi-objective dynamic optimization algorithm comprises:

[0059] S1. Initialize algorithm parameters, where the algorithm parameters include at least the number of particles in the population, the inertia weight of the particles, the initial velocity of the particles, and the initial position of the particles, wherein each particle in the population represents a feasible solution to the current power distribution task;

[0060] S2. Calculate the fitness value of each particle based on a preset fitness function;

[0061] S3. Determine the individual extreme value of each particle, the population extreme value and the average fitness value of the population according to the fitness value of each particle, wherein the individual extreme value is the optimal position searched by a single particle up to the present moment, and the population extreme value is the optimal position searched by the entire particle group up to the present moment;

[0062] S4, updating the inertia weight of each particle based on the average fitness value of the population, the maximum value of the inertia weight of the particle, and the minimum value of the inertia weight of the particle;

[0063] S5, update the current speed and current position of each particle according to the inertia weight, individual extreme value and population extreme value of each particle, return to step S2, if the population extreme value is not updated after N consecutive iterations, update the current speed and current position of each particle according to the individual extreme value and population extreme value;

[0064] S6. Determine whether the termination condition is met. If so, output the current optimal individual extreme value and the population extreme value, whichever has the larger fitness value, as the current optimal solution. Otherwise, return to step S2.

[0065] Optionally, the inertia weight of each particle is updated based on the average fitness value of the population, the maximum value of the inertia weight of the particle, and the minimum value of the inertia weight of the particle, including:

[0066] Update the inertia weight of each particle using the following formula:

[0067]

[0068] in, represents the inertia weight of the particle, Indicates the minimum value of the particle's inertia weight, Indicates the maximum value of the particle's inertia weight, represents the average fitness value of the population, g i ( t )express t Particle at time i Fitness value, min g i ( t ) represents the minimum fitness value of all particles.

[0069] Optionally, the current speed and current position of each particle are updated according to the individual extreme value and the population extreme value, including:

[0070] Update the current speed of each particle using the following formula:

[0071]

[0072] Update the current position of each particle using the following formula:

[0073]

[0074] in, Filial v represents the current velocity of the particle, Filial x represents the current position of the particle, c represents the learning factor, Pool v (seed i ) represents the speed of individual extreme values, Pool v (seed j ) represents the speed of the population extreme value, Pool x (seed i ) represents the location of individual extreme values, Pool x (seed j ) Indicates the location of the population extreme value, norm Indicates the modulus length.

[0075] Optionally, the mean field evolution equation includes:

[0076]

[0077] in, Represents the smart grid entity at time t and business status s The average field distribution under m ( t , s )express m ( t , s ) about the rate of change of energy state E, express t Energy usage status at all times, m ( t , s )express m ( t , s ) About queue status q The rate of change, express t The queue status at the moment.

[0078] Optionally, solving a preset mean field evolution equation based on a finite difference method includes:

[0079] Based on the finite difference method, the preset mean field evolution equation is solved and the following relationship is obtained:

[0080]

[0081] in, , , , Indicates time ,energy level and status The value of the mean field at Indicates time ,energy level and status The power at Indicates time ,energy level and status The value of the queue status at .

[0082] According to a third aspect of the present application, a machine-readable storage medium is provided, on which instructions are stored. When the instructions are executed by a processor, the processor is configured to execute the distributed intelligent power distribution optimization method as described above.

[0083] In a fourth aspect of the present application, a terminal device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned distributed intelligent power distribution optimization method when executing the computer program.

[0084] This application predicts the optimal distribution strategy of a dynamic heterogeneous network at the next moment based on mean field game, which can effectively reduce the complexity of optimizing the utility function of a large number of network entity nodes in a distributed intelligent distribution network, ensure the efficiency and reliability of the intelligent distribution network, and significantly improve the operating efficiency of the power grid, especially for large-scale and complex power grid systems.

[0085] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present application, but do not constitute a limitation on the embodiments of the present application. In the accompanying drawings:

[0087] Figure 1A method flow chart of a distributed intelligent power distribution optimization method provided in a preferred embodiment of the present application;

[0088] Figure 2 A flowchart of a multi-objective optimization algorithm provided for a preferred embodiment of the present application;

[0089] Figure 3 A schematic diagram of the optimal power distribution strategy calculation logic provided for the preferred embodiment of the present application;

[0090] Figure 4 A schematic block diagram of a distributed intelligent power distribution optimization device provided in a preferred embodiment of the present application;

[0091] Figure 5 A schematic diagram of a terminal device provided for a preferred embodiment of the present application.

[0092] Description of Reference Numerals

[0093] 10-terminal device, 100-processor, 101-memory, 102-computer program. DETAILED DESCRIPTION

[0094] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0095] It should be noted that the technical solutions between the various embodiments of the present application can be combined with each other, but it must be based on the fact that ordinary technicians in the field can implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0096] like Figure 1As shown, the first aspect of the present application provides a distributed intelligent power distribution optimization method, including: constructing a mean field distribution model that characterizes the probability distribution of different business states of each smart grid entity in the distributed smart grid; constructing a profit function of the smart grid entity based on the mean field distribution model, performing dynamic programming transformation on the profit function, and obtaining a value function that characterizes the mapping relationship between the profit and the distribution strategy of the smart grid entity under different business states and times, wherein the distribution strategy includes the business actions of the smart grid entity under the current business state and time; taking the value function as the control equation, solving the control equation based on a preset multi-objective dynamic optimization algorithm, and solving the preset mean field evolution equation based on the finite difference method, so as to obtain the optimal distribution strategy of each smart grid entity when both the control equation and the mean field evolution equation reach the Nash equilibrium point, and optimizing the distribution of each smart grid entity based on the optimal distribution strategy.

[0097] In this way, the present application predicts the optimal distribution strategy of a dynamic heterogeneous network at the next moment based on the mean field game, which can effectively reduce the complexity of optimizing and solving the utility function of a large number of network entity nodes in the distributed intelligent distribution network, ensure the efficiency and reliability of the intelligent distribution network, and significantly improve the operating efficiency of the power grid, especially for large-scale and complex power grid systems.

[0098] It is understandable that the smart grid entity can be a smart meter, distributed power source, energy storage device, etc. in the smart grid. The business state indicates the business situation of the smart grid entity at a certain moment. For example, the business state can be the remaining power, power demand, etc.; the business action indicates the action of the smart grid entity to select the business. For example, the business action can be discharging or charging the energy storage device, or changing the power consumption mode to reduce the power demand, etc. It is understandable that for smart meters, business actions include real-time monitoring of power usage, transmitting data to grid operators and consumers, and supporting demand response management; for distributed power sources, business actions include solar photovoltaic systems, wind power generation, small hydropower generation, and geothermal power generation facilities to generate electricity at the edge to reduce dependence on central power stations; for energy storage devices, business actions include energy storage, storing excess power for use during high demand, balancing grid loads, etc.; for grid automation equipment / Internet of Things equipment, business actions include real-time adjustment and optimization of grid operation; for communication equipment, business actions include ensuring real-time data communication between grid entities to achieve remote monitoring and dynamic control.

[0099] This application uses the mean field to represent the current business status of the smart grid entity and the mean field probability distribution, that is, the probability distribution of the status of all participants in the game. Among them, the mean field distribution model can be expressed as:

[0100]

[0101] in, , indicating the current business status of the smart grid entity, including the amount of electricity q and other energy states E , , represents the mean field distribution, I represents the number of smart grid entities, X is the state space of all possible states, Indicates that in the time interval [ 0 , T ] in all states. Among them, Also depends on the state and time , Represents an indicator function that returns 1 if the condition is true, otherwise it returns 0. The mean field distribution model is used to predict the evolution of the average behavior or state in a large number of grid entities.

[0102] In this application, the revenue function of the smart grid entity is constructed based on the mean field distribution model, including the following steps:

[0103] According to the preset business revenue factor, average field impact factor, transmission cost factor, computing cost factor, communication transmission resources and computing resources of smart grid entities, a cost-benefit function corresponding to the business status, business action and average field distribution of smart grid entities is constructed. Among them, the cost-benefit function can be expressed as:

[0104]

[0105] in, represents the business profit factor, represents the mean field influence factor, represents the transmission cost factor, Indicates communication transmission resources, Indicates the business action selected by the smart grid entity, represents the computational cost factor, represents the computing resources allocated to the smart grid entity, m represents the mean field distribution, and x represents the business status of the smart grid entity.

[0106] According to the preset penalty factor, the total amount of business tasks corresponding to the current power distribution task, the current business status of each smart grid entity and the number of smart grid entities, a penalty function is constructed to represent the terminal penalty of each smart grid entity under different business task allocation states. The penalty function can be expressed as:

[0107]

[0108] in, represents the penalty factor, Indicates the total amount of business tasks. Indicates The business status of each smart grid entity, Indicates the number of smart grid entities within the scope. The total business task volume refers to the sum of all power demands that need to be distributed and managed by the smart grid system during a specific period of time, including power demands under various business scenarios such as residential power, commercial power, and industrial power, such as power distribution, energy storage, demand response management, and other tasks. The total business task volume can be derived based on historical data, real-time data, prediction models, and other information, and is not limited here.

[0109] The profit function of the smart grid entity is constructed based on the cost-benefit function and the penalty function, where the profit function can be expressed as:

[0110]

[0111] in, represents the performance indicator function, which represents the benefit of this process, Indicates the business action selected by the smart grid entity, is the cost-benefit function, which represents the cost-benefit generated in this process. Indicates the current business status of the smart grid entity, yes The vector formed can be expressed as , is a penalty function. In this application, if there is any unallocated service, the current smart grid entity will be penalized.

[0112] In this application, a dynamic programming transformation is performed on the revenue function to obtain a value function that characterizes the mapping relationship between revenue and distribution strategy of the smart grid entity under different business states and time, including:

[0113] Based on the Bellman optimality principle, the obtained revenue function is converted into the value function of the smart grid entity, where the obtained Bellman value function can be expressed as:

[0114]

[0115] After obtaining the value function representing the mapping relationship between the revenue and the distribution strategy of the smart grid entity under different business states and times, the method further includes: converting the value function based on a preset time variation, performing Taylor expansion on the converted value function to obtain a corresponding objective function, specifically, taking a very small time variation , transform the value function into the following form: , Taylor expansion is performed on the formula, and the corresponding objective function is obtained as follows: , let the time variation tend to a preset value, in this application, the preset value is 0, so as to generate the HJB equation representing the power distribution strategy and business status of the smart grid entity based on the objective function. That is, when dt tends to zero, the partial differential equation of the power distribution strategy and business status adopted by the smart grid entity can be obtained from the objective function as follows:

[0116]

[0117] in, o ( dt ) is a high-order infinitesimal, indicating that as dt The part that tends to zero tends to zero, and They are respectively Bellman value function with respect to business status x and time t The partial derivative of α ( x ( t )) indicates the time t Business status x The optimal business action to be performed under.

[0118] In this application, the value function is used as the control equation, that is, the HJB equation is used as the control equation.

[0119] like Figure 2 As shown, the multi-objective dynamic optimization algorithm of the present application includes:

[0120] S1. Initialize the algorithm parameters. The algorithm parameters include at least the number of particles in the population, the inertia weight of the particles, the initial velocity of the particles, and the initial position of the particles. Each particle in the population represents a feasible solution to the current power distribution task, that is, the representation of the business actions of each smart grid entity. Assuming that each smart grid entity has a service task, each particle individual is represented by a vector in the algorithm. Each dimension of the vector represents the task allocation strategy of a smart grid entity, that is, one dimension of the vector represents the business action of a smart grid entity. It can be understood that in the smart grid environment, each smart grid entity is responsible for a service task. For example, the service tasks include demand response management, distributed grid management, energy storage management, power supply restoration and other operations.

[0121] S2. Calculate the fitness value of each particle based on a preset fitness function. The fitness function is:

[0122]

[0123] in, L Represents the total number of levels / dimensions, specifically the total number of grid entities or service types that need to be optimized. DRepresents the set of all grid entities, each of which needs to be optimized. K It represents the set of possible strategies that each grid entity may adopt. In each iteration, each entity can select a strategy from this set. Indicated in t In the iteration, d The grid entity takes k performance indicators when using a strategy.

[0124] S3. Determine the individual extreme value of each particle, the population extreme value and the average fitness value of the population based on the fitness value of each particle, wherein the individual extreme value is the optimal position searched by a single particle up to the present moment, and the population extreme value is the optimal position searched by the entire particle group up to the present moment.

[0125] S4. Update the inertia weight of each particle based on the average fitness value of the population, the maximum value of the inertia weight of the particle, and the minimum value of the inertia weight of the particle. In this application, the inertia weight of each particle is updated by the following formula:

[0126]

[0127] in, represents the inertia weight of the particle, Indicates the minimum value of the particle's inertia weight, Indicates the maximum value of the particle's inertia weight, represents the average fitness value of the population, g i ( t )express t Particle at time i Fitness value, min g i ( t ) represents the minimum fitness value of all particles. The purpose of the weight adjustment method of the present application is to adjust the search behavior of the particle according to the difference between the current fitness of the particle and the optimal fitness, so that the particle increases or decreases its inertia according to its performance relative to the performance of the group, thereby optimizing the balance between global and local searches, so as to achieve the best performance at different optimization stages. S5, update the current speed and current position of each particle according to the inertia weight, individual extreme value and population extreme value of each particle, return to step S2, if the population extreme value is not updated after N consecutive iterations, update the current speed and current position of each particle according to the individual extreme value and the population extreme value. It can be understood that the updating of the current speed and current position of each particle according to the inertia weight, individual extreme value and population extreme value of each particle can be implemented by the existing technology, which is not limited here.

[0128] Among them, this application updates the current speed of each particle through the following formula:

[0129]

[0130] Update the current position of each particle using the following formula:

[0131]

[0132] in, Filial v represents the current velocity of the particle, Filial x represents the current position of the particle, c represents the learning factor, Pool v (seed i ) represents the speed of individual extreme values, Pool v (seed j ) represents the speed of the population extreme value, Pool x (seed i ) represents the location of individual extreme values, Pool x (seed j ) Indicates the location of the population extreme value, norm Indicates the modulus length. It can be understood that the individual extreme value is the individual best historical position, that is, the best solution found by the particle itself, the population group extreme value is the group's best historical position, that is, the best solution among all particles, and the learning factor is used to adjust the relative influence factor of individual learning and group learning. The present application uses the modulus length of the vector to adjust the proportion of speed update so that the direction of the new speed is consistent with the direction of the individual best and the group best, but its size is affected by the ratio of the modulus lengths in the two directions. The method of the present application can help particles explore more effectively in the search space while maintaining focus on high-quality areas and improving search efficiency.

[0133] S6. Determine whether the termination condition is met. If so, output the current optimal individual extreme value and the population group extreme value, whichever has the larger fitness value, as the current optimal solution. Otherwise, return to step S2. It can be understood that the termination condition may be reaching the maximum number of iterations. The present application determines the optimal power distribution strategy by integrating the improved particle swarm algorithm with the mean field evolution equation. For the calculation process of the optimal power distribution strategy, the termination condition may be reaching the maximum number of iterations, and the control equation and the mean field evolution equation reach the Nash equilibrium point.

[0134] Through the multi-objective dynamic optimization algorithm of the present application, the ordinary particle swarm algorithm can jump out of the local optimal solution when facing the local optimum, thereby achieving more effective multi-objective dynamic optimization, which helps to improve the global search capability of the algorithm and enables it to better cope with complex dynamic environments and multi-objective optimization problems.

[0135] like Figure 3 As shown, the present application introduces the mean field evolution equation and utilizes the data of other entities in the network to predict and adjust local strategies, thereby effectively reducing the communication load because each entity does not have to continuously monitor all other entities in the network, but instead adjusts the local output based on the mean field data.

[0136] Since the actions of each participant in the game process depend on the decisions taken by other participants, the HJB equation derived by each smart grid entity contains vectors composed of the decisions taken by other participants. These decisions will be coupled and affect each other. At the same time, the HJB equations will also be coupled with each other. In order to obtain the mean field equilibrium, this application uses the finite difference method to convert the time interval , energy usage state space , queue state space Discrete Space, that is, exists separately time point, Status and The iteration steps are defined as: , , That is, the evolution space of the FPK equation is ,At the same time, by solving the HJB equation in reverse, the optimal path can be obtained.

[0137] Specifically, the mean field evolution equation of the present application, namely the FPK equation, is as follows:

[0138]

[0139] in, Represents the smart grid entity at time t and business status s The average field distribution under the condition can be understood as the distribution of power grid entities or the state probability at a specific time and state; m ( t , s )express m ( t , s ) Regarding the rate of change of energy state E, this distribution can be understood as the distribution or state probability of power grid entities at a specific time and state; expresst Energy usage status at all times, m ( t , s )express m ( t , s ) About queue status q The rate of change of shows how the queue status affects the status distribution of the entire power grid; express t The energy state and queue state represent the specific parameters of smart grid entities, which are used to describe and analyze the behavior and performance of smart grids at different times and under different conditions, so as to optimize the operation of the grid and improve efficiency and reliability through precise adjustment and prediction of the state.

[0140] Wherein, solving the preset mean field evolution equation based on the finite difference method, such as the Lax-Friedrichs method, includes: solving the preset mean field evolution equation based on the finite difference method to obtain the following relationship:

[0141]

[0142] in, , , , Indicates time ,energy level and status The value of the mean field at Indicates time ,energy level and status The power at Indicates time ,energy level and status The value of the queue status at . M ( i+1 , j , k ) represents the state distribution of the intelligent entity at the next time step. M Represents the state distribution of the current time step, which is associated with the actual state in the power grid, such as power load, equipment operating status, etc.; P and Q These two parameters play a regulating role in the equation. P Represents energy-related regulatory factors or parameters, such as distribution strategies in a power grid or optimization measures for power flow, which involve how to allocate power resources to meet demand or maximize efficiency. Q It is a regulation factor or parameter related to the queue, such as the task status or task queue in the power grid, such as the power demand task waiting to be processed, the maintenance task, etc. This application combines the particle swarm speed update formula

[0143] , position update formula With the mean field update formula , update the particle's inertia weight according to the particle's fitness and the state of the mean field ,Right now . Then determine the stopping condition. If the stopping condition is met, such as reaching the maximum number of iterations or the global optimal position, output the current optimal solution; otherwise, continue iterating. This application effectively integrates the above-mentioned particle swarm optimization algorithm with the HJB equation and FPK equation in the mean field game, thereby improving the algorithm's global search capability and optimization effect when dealing with complex dynamic environments and multi-objective optimization problems.

[0144] For example, the goal is to optimize the distribution strategy in the urban power grid to reduce overall energy consumption and improve the stability of the power grid. The specific steps of the mean field game and particle swarm optimization based on this application are as follows:

[0145] 1. Minimize the total energy consumption of the power grid while ensuring the load balance of the power grid. It is understandable that each smart grid entity such as substation, distribution point, distributed power source, etc. has different load requirements and energy usage.

[0146] 2. Initialize the particle swarm: Set each particle to represent a possible power distribution strategy, and the particle position (strategy vector) represents the energy distribution of each grid entity. The initial position can be a randomly generated energy distribution strategy, and the initial speed is set to zero or a small random value.

[0147] 3. Calculate fitness: Calculate the total energy consumption and load balance according to the strategy of each particle through the fitness function.

[0148] 4. Update strategy:

[0149] 1) Particle swarm optimization updates individual extreme values ​​and group extreme values, and then updates the speed and position.

[0150] 2) Update the mean field evolution equation, use the finite difference method to update the mean field distribution, calculate the strategy adjustment of each grid entity, and reduce the load.

[0151] 5. Iteration and optimization: Check whether the maximum number of iterations has been reached. If not, continue to iterate steps 3 to 5. If reached, output the current optimal power distribution strategy.

[0152] It can be understood that the mean field evolution equation gives the evolution process of the energy distribution of the smart grid entity. After multiple iterations, the distribution strategy and the mean field of the smart grid entity no longer change. At this time, the Nash equilibrium point is reached, and the distribution decision vector that reaches the Nash equilibrium point is determined to be the optimal decision vector.

[0153] This application can adjust resource allocation in real time and optimize energy consumption; at the same time, by analyzing the energy consumption and performance data of different power grid entities, the system can prioritize the allocation of more efficient and lower-cost resources to maximize cost-effectiveness.

[0154] like Figure 4 As shown, in a second aspect of the present application, a distributed intelligent power distribution optimization device is provided, comprising:

[0155] A model building module is configured to build a mean field distribution model representing the probability distribution of different business states of each smart grid entity in the distributed smart grid;

[0156] A revenue function construction module is configured to construct a revenue function of a smart grid entity based on a mean field distribution model, perform dynamic programming transformation on the revenue function, and obtain a value function representing a mapping relationship between revenue and power distribution strategy of the smart grid entity under different business states and times, wherein the power distribution strategy includes business actions of the smart grid entity under current business states and times;

[0157] The power distribution optimization module is configured to use the value function as the control equation, solve the control equation based on a preset multi-objective dynamic optimization algorithm, and solve the preset mean field evolution equation based on the finite difference method, so as to obtain the optimal power distribution strategy of each smart grid entity when both the control equation and the mean field evolution equation reach the Nash equilibrium point, and optimize the power distribution of each smart grid entity based on the optimal power distribution strategy.

[0158] Optionally, a revenue function of a smart grid entity is constructed based on the mean field distribution model, including:

[0159] According to the preset business revenue factor, average field impact factor, transmission cost factor, computing cost factor, communication transmission resources and computing resources of smart grid entities, a cost-benefit function corresponding to the business status, business action and average field distribution of smart grid entities is constructed;

[0160] According to the preset penalty factor, the total amount of business tasks corresponding to the current power distribution task, the current business status of each smart grid entity and the number of smart grid entities, a penalty function representing the terminal penalty of each smart grid entity under different business task allocation states is constructed;

[0161] The profit function of smart grid entities is constructed based on the cost-benefit function and penalty function.

[0162] Optionally, a dynamic programming transformation is performed on the revenue function to obtain a value function that characterizes the mapping relationship between the revenue and the power distribution strategy of the smart grid entity under different business states and times, including:

[0163] Based on the Bellman optimality principle, the revenue function is converted into the value function of the smart grid entity;

[0164] After obtaining the value function that characterizes the mapping relationship between the revenue and the distribution strategy of the smart grid entity under different business states and time, the revenue function construction module is further configured as follows:

[0165] The value function is converted based on a preset time variation, and the converted value function is Taylor expanded to obtain a corresponding objective function;

[0166] The time variation is made to approach a preset value so as to generate the HJB equation representing the power distribution strategy and business status of the smart grid entity based on the objective function.

[0167] Optionally, the value function is used as the control equation, including:

[0168] The HJB equation is used as the control equation;

[0169] Multi-objective dynamic optimization algorithm, including:

[0170] S1. Initialize algorithm parameters, which include at least the number of particles in the population, the inertia weight of the particles, the initial velocity of the particles, and the initial position of the particles. Each particle in the population represents a feasible solution to the current power distribution task.

[0171] S2. Calculate the fitness value of each particle based on a preset fitness function;

[0172] S3. Determine the individual extreme value of each particle, the population extreme value and the average fitness value of the population according to the fitness value of each particle, wherein the individual extreme value is the optimal position searched by a single particle up to the present moment, and the population extreme value is the optimal position searched by the entire particle group up to the present moment;

[0173] S4, updating the inertia weight of each particle based on the average fitness value of the population, the maximum value of the inertia weight of the particle, and the minimum value of the inertia weight of the particle;

[0174] S5, update the current speed and current position of each particle according to the inertia weight, individual extreme value and population extreme value of each particle, return to step S2, if the population extreme value is not updated after N consecutive iterations, update the current speed and current position of each particle according to the individual extreme value and population extreme value;

[0175] S6. Determine whether the termination condition is met. If so, output the current optimal individual extreme value and the population extreme value, whichever has the larger fitness value, as the current optimal solution. Otherwise, return to step S2.

[0176] Optionally, the inertia weight of each particle is updated based on the average fitness value of the population, the maximum value of the inertia weight of the particle, and the minimum value of the inertia weight of the particle, including:

[0177] Update the inertia weight of each particle using the following formula:

[0178]

[0179] in, represents the inertia weight of the particle, Indicates the minimum value of the particle's inertia weight, Indicates the maximum value of the particle's inertia weight, represents the average fitness value of the population, g i ( t )express t Particle at time i Fitness value, min g i ( t ) represents the minimum fitness value of all particles.

[0180] Optionally, the current speed and current position of each particle are updated according to the individual extreme value and the population extreme value, including:

[0181] Update the current speed of each particle using the following formula:

[0182]

[0183] Update the current position of each particle using the following formula:

[0184]

[0185] in, Filial v represents the current velocity of the particle, Filial x represents the current position of the particle, c represents the learning factor, Pool v (seedi ) represents the speed of individual extreme values, Pool v (seed j ) represents the speed of the population extreme value, Pool x (seed i ) represents the location of individual extreme values, Pool x (seed j ) Indicates the location of the population extreme value, norm Indicates the modulus length.

[0186] Optionally, the mean field evolution equations include:

[0187]

[0188] in, Represents the smart grid entity at time t and business status s The average field distribution under m ( t , s )express m ( t , s ) about the rate of change of energy state E, express t Energy usage status at all times, m ( t , s )express m ( t , s ) About queue status q The rate of change, express t The queue status at the moment.

[0189] Optionally, solving a preset mean field evolution equation based on a finite difference method includes:

[0190] Based on the finite difference method, the preset mean field evolution equation is solved and the following relationship is obtained:

[0191]

[0192] in, , , , Indicates time ,energy level and status The value of the mean field at Indicates time ,energy level and status The power at Indicates time ,energy level and status The value of the queue status at .

[0193] It is understandable that the technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0194] According to a third aspect of the present application, a machine-readable storage medium is provided, on which instructions are stored. When the instructions are executed by a processor, the processor is configured to execute the distributed intelligent power distribution optimization method as described above.

[0195] In a fourth aspect of the present application, a terminal device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned distributed intelligent power distribution optimization method when executing the computer program.

[0196] like Figure 5 FIG. 1 is a schematic diagram of a terminal device provided in an embodiment of the present application. Figure 5 As shown, the terminal device 10 of this embodiment includes: a processor 100, a memory 101, and a computer program 102 stored in the memory 101 and executable on the processor 100. When the processor 100 executes the computer program 102, the steps in the above method embodiment are implemented. Alternatively, when the processor 100 executes the computer program 102, the functions of each module / unit in the above device embodiments are implemented.

[0197] Exemplarily, the computer program 102 may be divided into one or more modules / units, one or more modules / units are stored in the memory 101, and are executed by the processor 100 to complete the present application. One or more modules / units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 102 in the terminal device 10.

[0198] The terminal device 10 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The terminal device 10 may include, but is not limited to, a processor 100 and a memory 101. Those skilled in the art will appreciate that Figure 5 It is only an example of the terminal device 10 and does not constitute a limitation of the terminal device 10. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device may also include input and output devices, network access devices, buses, etc.

[0199] The processor 100 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0200] The memory 101 may be an internal storage unit of the terminal device 10, such as a hard disk or memory of the terminal device 10. The memory 101 may also be an external storage device of the terminal device 10, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card (FlashCard), etc. equipped on the terminal device 10. Further, the memory 101 may also include both an internal storage unit of the terminal device 10 and an external storage device. The memory 101 is used to store computer programs and other programs and data required by the terminal device 10. The memory 101 may also be used to temporarily store data that has been output or is to be output.

[0201] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0202] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0203] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A distributed intelligent power distribution optimization method, characterized in that: include: Construct a mean field distribution model to characterize the probability distribution of different service states of each smart grid entity in a distributed smart grid; Constructing a revenue function of a smart grid entity based on the mean field distribution model, performing dynamic programming transformation on the revenue function, and obtaining a value function that characterizes a mapping relationship between revenue and power distribution strategy of the smart grid entity under different business states and times, wherein the power distribution strategy includes business actions of the smart grid entity under the current business state and time; Taking the value function as the control equation, solving the control equation based on a preset multi-objective dynamic optimization algorithm, and solving the preset mean field evolution equation based on a finite difference method, so as to obtain the optimal power distribution strategy of each smart grid entity when both the control equation and the mean field evolution equation reach a Nash equilibrium point, and optimizing the power distribution of each smart grid entity based on the optimal power distribution strategy; The multi-objective dynamic optimization algorithm comprises: S1. Initialize algorithm parameters, where the algorithm parameters include at least the number of particles in the population, the inertia weight of the particles, the initial velocity of the particles, and the initial position of the particles, wherein each particle in the population represents a feasible solution to the current power distribution task; S2. Calculate the fitness value of each particle based on a preset fitness function; Among them, the fitness function is: in, L represents the total number of smart grid entities, D represents the set of all smart grid entities, K represents the set of policies for each smart grid entity, Indicated in t In the iteration, d Smart grid entities take the k Performance indicators when using a strategy; S3. Determine the individual extreme value of each particle, the population extreme value and the average fitness value of the population according to the fitness value of each particle, wherein the individual extreme value is the optimal position searched by a single particle up to the present moment, and the population extreme value is the optimal position searched by the entire particle group up to the present moment; S4. Update the inertia weight of each particle using the following formula: in, represents the inertia weight of the particle, Indicates the minimum value of the particle's inertia weight, Indicates the maximum value of the particle's inertia weight, represents the average fitness value of the population, g i ( t )express t Particle at time i Fitness value, min g i ( t ) represents the minimum fitness value of all particles; S5, update the current speed and current position of each particle according to the inertia weight, individual extreme value and population extreme value of each particle, return to step S2, if the population extreme value is not updated after N consecutive iterations, update the current speed and current position of each particle according to the individual extreme value and population extreme value; S6. Determine whether the termination condition is met. If so, output the current optimal individual extreme value and the population extreme value, whichever has the larger fitness value, as the current optimal solution. Otherwise, return to step S2.

2. The distributed intelligent power distribution optimization method according to claim 1, characterized in that: The revenue function of the smart grid entity is constructed based on the average field distribution model, including: According to the preset business revenue factor, average field impact factor, transmission cost factor, computing cost factor, communication transmission resources and computing resources of smart grid entities, a cost-benefit function corresponding to the business status, business action and average field distribution of smart grid entities is constructed; According to the preset penalty factor, the total amount of business tasks corresponding to the current power distribution task, the current business status of each smart grid entity and the number of smart grid entities, a penalty function representing the terminal penalty of each smart grid entity under different business task allocation states is constructed; A benefit function of a smart grid entity is constructed based on the cost benefit function and the penalty function.

3. The distributed intelligent power distribution optimization method according to claim 1, characterized in that: Performing dynamic programming transformation on the revenue function to obtain a value function that characterizes the mapping relationship between revenue and distribution strategy of the smart grid entity under different business states and times, including: Based on the Bellman optimality principle, the revenue function is converted into a value function of a smart grid entity; After obtaining the value function representing the mapping relationship between the revenue and the power distribution strategy of the smart grid entity under different business states and times, the method further includes: Converting the value function based on a preset time variation, and performing Taylor expansion on the converted value function to obtain a corresponding objective function; The time variation is made to tend to a preset value, so as to generate a HJB equation representing the power distribution strategy and business status of the smart grid entity based on the objective function.

4. The distributed intelligent power distribution optimization method according to claim 3, characterized in that: The value function is used as the control equation, including: The HJB equation is used as the control equation.

5. The distributed intelligent power distribution optimization method according to claim 4, characterized in that: Update the current speed and position of each particle based on individual extreme values ​​and population extreme values, including: Update the current speed of each particle using the following formula: Update the current position of each particle using the following formula: in, Filial v represents the current velocity of the particle, Filial x represents the current position of the particle, c represents the learning factor, Pool v (seed i ) represents the speed of individual extreme values, Pool v (seed j ) represents the speed of the population extreme value, Pool x (seed i ) represents the location of individual extreme values, Pool x (seed j ) Indicates the location of the population extreme value, norm Indicates the modulus length.

6. The distributed intelligent power distribution optimization method according to claim 1, characterized in that: The mean field evolution equation includes: in, Represents the smart grid entity at time t and business status s The average field distribution under m ( t , s )express m ( t , s ) about the rate of change of energy state E, express t Energy usage status at all times, m ( t , s )express m ( t , s ) About queue status q The rate of change, express t The queue status at the moment.

7. The distributed intelligent power distribution optimization method according to claim 6, characterized in that: The preset mean field evolution equations are solved based on the finite difference method, including: Based on the finite difference method, the preset mean field evolution equation is solved and the following relationship is obtained: in, , , , Indicates time ,energy level and status The value of the mean field at Indicates time ,energy level and status The power at Indicates time ,energy level and status The value of the queue status at .

8. A distributed intelligent power distribution optimization device, characterized in that: include: A model building module is configured to build a mean field distribution model representing the probability distribution of different business states of each smart grid entity in the distributed smart grid; A revenue function construction module is configured to construct a revenue function of a smart grid entity based on the mean field distribution model, perform dynamic programming transformation on the revenue function, and obtain a value function that characterizes a mapping relationship between revenue and power distribution strategy of the smart grid entity under different business states and times, wherein the power distribution strategy includes a business action of the smart grid entity under a current business state and time; A power distribution optimization module is configured to use the value function as a control equation, solve the control equation based on a preset multi-objective dynamic optimization algorithm, and solve a preset mean field evolution equation based on a finite difference method, so as to obtain an optimal power distribution strategy for each smart grid entity when both the control equation and the mean field evolution equation reach a Nash equilibrium point, and perform power distribution optimization for each smart grid entity based on the optimal power distribution strategy; The multi-objective dynamic optimization algorithm comprises: S1. Initialize algorithm parameters, where the algorithm parameters include at least the number of particles in the population, the inertia weight of the particles, the initial velocity of the particles, and the initial position of the particles, wherein each particle in the population represents a feasible solution to the current power distribution task; S2. Calculate the fitness value of each particle based on a preset fitness function; Among them, the fitness function is: in, L represents the total number of smart grid entities, D represents the set of all smart grid entities, K represents the set of policies for each smart grid entity, Indicated in t In the iteration, d Smart grid entities take the k Performance indicators when using a strategy; S3. Determine the individual extreme value of each particle, the population extreme value and the average fitness value of the population according to the fitness value of each particle, wherein the individual extreme value is the optimal position searched by a single particle up to the present moment, and the population extreme value is the optimal position searched by the entire particle group up to the present moment; S4. Update the inertia weight of each particle using the following formula: in, represents the inertia weight of the particle, Indicates the minimum value of the particle's inertia weight, Indicates the maximum value of the particle's inertia weight, represents the average fitness value of the population, g i ( t )express t Particle at time i Fitness value, min g i ( t ) represents the minimum fitness value of all particles; S5, update the current speed and current position of each particle according to the inertia weight, individual extreme value and population extreme value of each particle, return to step S2, if the population extreme value is not updated after N consecutive iterations, update the current speed and current position of each particle according to the individual extreme value and population extreme value; S6. Determine whether the termination condition is met. If so, output the current optimal individual extreme value and the population extreme value, whichever has the larger fitness value, as the current optimal solution. Otherwise, return to step S2.

9. The distributed intelligent power distribution optimization device according to claim 8, characterized in that: The revenue function of the smart grid entity is constructed based on the average field distribution model, including: According to the preset business revenue factor, average field impact factor, transmission cost factor, computing cost factor, communication transmission resources and computing resources of smart grid entities, a cost-benefit function corresponding to the business status, business action and average field distribution of smart grid entities is constructed; According to the preset penalty factor, the total amount of business tasks corresponding to the current power distribution task, the current business status of each smart grid entity and the number of smart grid entities, a penalty function representing the terminal penalty of each smart grid entity under different business task allocation states is constructed; A benefit function of a smart grid entity is constructed based on the cost benefit function and the penalty function.

10. The distributed intelligent power distribution optimization device according to claim 9, characterized in that: Performing dynamic programming transformation on the revenue function to obtain a value function that characterizes the mapping relationship between revenue and distribution strategy of the smart grid entity under different business states and times, including: Based on the Bellman optimality principle, the revenue function is converted into a value function of a smart grid entity; After obtaining the value function representing the mapping relationship between the revenue and the power distribution strategy of the smart grid entity under different business states and times, the revenue function construction module is further configured as follows: Converting the value function based on a preset time variation, and performing Taylor expansion on the converted value function to obtain a corresponding objective function; The time variation is made to tend to a preset value, so as to generate a HJB equation representing the power distribution strategy and business status of the smart grid entity based on the objective function.

11. The distributed intelligent power distribution optimization device according to claim 10, characterized in that: The value function is used as the control equation, including: The HJB equation is used as the control equation.

12. The distributed intelligent power distribution optimization device according to claim 11, characterized in that: Update the current speed and position of each particle based on individual extreme values ​​and population extreme values, including: Update the current speed of each particle using the following formula: Update the current position of each particle using the following formula: in, Filial v represents the current velocity of the particle, Filial x represents the current position of the particle, c represents the learning factor, Pool v (seed i ) represents the speed of individual extreme values, Pool v (seed j ) represents the speed of the population extreme value, Pool x (seed i ) represents the location of individual extreme values, Pool x (seed j ) Indicates the location of the population extreme value, norm Indicates the modulus length.

13. The distributed intelligent power distribution optimization device according to claim 8, characterized in that: The mean field evolution equation includes: in, Represents the smart grid entity at time t and business status s The average field distribution under m ( t , s )express m ( t , s ) about the rate of change of energy state E, express t Energy usage status at all times, m ( t , s )express m ( t , s ) About queue status q The rate of change, express t The queue status at the moment.

14. The distributed intelligent power distribution optimization device according to claim 13, characterized in that: The preset mean field evolution equations are solved based on the finite difference method, including: Based on the finite difference method, the preset mean field evolution equation is solved and the following relationship is obtained: in, , , , Indicates time ,energy level and status The value of the mean field at Indicates time ,energy level and status The power at Indicates time ,energy level and status The value of the queue status at .

15. A machine-readable storage medium having instructions stored thereon, characterized in that: When the instruction is executed by a processor, the processor is configured to execute the distributed intelligent power distribution optimization method as described in any one of claims 1 to 7.

16. A terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the distributed intelligent power distribution optimization method as described in any one of claims 1 to 7 are implemented.

Images