Multi-agent cooperative game method and system for integrated energy system

Abstract

The invention discloses a multi-agent cooperative game method and system for an integrated energy system. The method comprises the following steps: constructing a cogeneration park model; the upper layer takes the power and gas network as a main body, takes the comprehensive demand response price as a decision variable, and takes profit maximization of the power and gas network as an optimization target; the lower layer takes photovoltaic, wind power and various parks as main bodies, considers the loss cost and carbon quota in the energy transmission process among the main bodies, takes the equipment output of each main body as a decision variable, and takes the maximum benefit of the multi-main body alliance as an optimization target to establish an operation optimization model of the double-layer game alliance; solving the operation optimization model of the double-layer game alliance by adopting an improved way finder algorithm to obtain the maximum multi-subject alliance income; and distributing the multi-agent alliance income to obtain the benefit distribution of the multiple agents under the condition of the strongest willingness to participate in the alliance. According to the invention, the renewable energy consumption capability in the park is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of integrated energy systems, and in particular to a multi-agent cooperative game method and system for integrated energy systems. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] The construction of an integrated energy system can effectively improve energy efficiency and reduce carbon emissions, which is of great significance to the sustainable development of energy. At the same time, hydrogen energy has the characteristics of high utilization rate and cleanness and no pollution, and has great potential in achieving low carbon.

[0004] As the electricity market becomes increasingly complex, cooperative games have been widely used in multi-agent decision-making optimization of integrated energy systems. Cooperative games can balance the demands of various stakeholders and achieve optimal scheduling of demand responses for multiple heterogeneous energy sources. Combining the three can further improve the overall energy utilization efficiency, reduce system operating costs, and reduce carbon emissions.

[0005] However, when multiple stakeholders participate in a cooperative game alliance at the same time, due to the complex economic behaviors among the stakeholders and the coupled energy equipment, it is difficult for the stakeholders to form a large alliance using conventional interest distribution methods, which weakens the energy interaction between the original multiple stakeholders, thereby increasing the operating costs and carbon emissions of each stakeholder, and reducing the willingness of each stakeholder to form a cooperative alliance. Summary of the invention

[0006] In order to solve the above problems, the present invention proposes a multi-agent cooperative game method and system for an integrated energy system, constructs a two-layer game model, and solves the optimal solution of the two-layer game model by iteratively optimizing the comprehensive demand response price at the upper layer and the output of the energy conversion coupling equipment in each subject at the lower layer, so as to achieve the highest profit of the electricity and gas grids and the overall profit of the alliance.

[0007] In some embodiments, the following technical solutions are adopted:

[0008] A multi-agent cooperative game method for an integrated energy system, comprising:

[0009] Based on the operating characteristics of the main equipment in the integrated energy system, a combined heat and power park model representing urban suburbs, urban industry, and urban commerce is constructed;

[0010] The upper layer takes the electricity and gas grids as the main body, the comprehensive demand response price as the decision variable, and the maximization of the profits of the electricity and gas grids as the optimization goal;

[0011] The lower layer takes photovoltaic, wind power, and various parks as the main bodies, considers the loss costs and carbon quotas in the energy transmission process between the main bodies, takes the equipment output of each main body as the decision variable, takes the maximum benefit of the multi-main body alliance as the optimization goal, and establishes the operation optimization model of the two-layer game alliance;

[0012] The improved Pathfinder algorithm is used to solve the operation optimization model of the two-layer game alliance and obtain the maximum multi-agent alliance benefit;

[0013] The benefits of the multi-agent alliance are distributed using the benefit distribution method, and the benefit distribution is obtained when the multi-agents have the strongest willingness to participate in the alliance.

[0014] Among them, the improved pathfinder algorithm is used to solve the operation optimization model of the two-layer game alliance. The specific process is as follows:

[0015] Initialize the parameters of the pathfinder algorithm; the initialized parameters include the random vector of the group, the maximum number of iterations, the fluctuation vector and the control parameter; the above-layer comprehensive demand response price is taken as the individual;

[0016] Calculate the upper layer objective function for each individual in the initialized population, find the current global optimal solution, set the optimal individual in the initialization and its objective function value as the pathfinder, and store them in the Archive set;

[0017] The two control parameters of the pathfinder algorithm are adaptively selected by the fuzzy control method; the wave vector is calculated according to the Levy flight method;

[0018] Calculate the optimal individual of the Pathfinder algorithm population and its objective function value, and update the location information of each follower; compare the new optimal individual with each individual in the Archive set to determine whether it is stored in the Archive set until the number of individuals in the Archive set reaches the upper limit; output the Archive set, calculate the cooperative game alliance profit under each individual in the Archive set, and select the individual that maximizes the cooperative game alliance profit and the corresponding equipment output and optimal profit.

[0019] In other embodiments, the following technical solutions are adopted:

[0020] A multi-agent cooperative game system for an integrated energy system, comprising:

[0021] The integrated energy system model building module is configured to build a combined heat and power park model representing urban suburbs, urban industries, and urban commerce based on the operating characteristics of the main equipment in the integrated energy system;

[0022] The operation optimization model building module is configured as follows: the upper layer takes the electricity and gas grids as the main body, takes the comprehensive demand response price as the decision variable, and takes the maximization of the profit of the electricity and gas grids as the optimization goal; the lower layer takes photovoltaic, wind power, and various parks as the main body, considers the loss cost and carbon quota in the energy transmission process between the main bodies, takes the equipment output of each main body as the decision variable, takes the maximization of the multi-body alliance profit as the optimization goal, and establishes the operation optimization model of the two-layer game alliance;

[0023] The model solving module is configured to use the improved pathfinder algorithm to solve the operation optimization model of the two-layer game alliance to obtain the maximum multi-agent alliance benefit;

[0024] The benefit distribution module is configured to use the benefit distribution method to distribute the benefits of the multi-agent alliance, and obtain the benefit distribution when the multi-agents have the strongest willingness to participate in the alliance.

[0025] In other embodiments, the following technical solutions are adopted:

[0026] A terminal device comprises a processor and a memory, wherein the processor is used to implement instructions; the memory is used to store a plurality of instructions, wherein the instructions are suitable for being loaded by the processor and executing the above-mentioned multi-agent cooperative game method of the integrated energy system.

[0027] In other embodiments, the following technical solutions are adopted:

[0028] A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor of a terminal device and executing the above-mentioned multi-agent cooperative game method for an integrated energy system.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] (1) The present invention constructs an operation optimization model of a two-layer game alliance. The upper layer is based on the electricity and gas grids, adopts time-of-day electricity and gas prices, takes the comprehensive demand response price as the decision variable, and takes the profit maximization of the electricity and gas grids as the objective function; the lower layer is based on photovoltaic, wind power, and various parks, takes the equipment output of each stakeholder as the decision variable, and constructs the overall profit objective function of the alliance in view of the different geographical locations of each subject. In the cooperative game alliance, energy interaction generates energy loss and carbon emissions due to the large distance, and the overall profit objective function of the alliance is constructed; the comprehensive demand response price of the upper layer and the output of the energy conversion coupling equipment in each subject of the lower layer are iteratively optimized to solve the optimal solution of the operation optimization model of the two-layer game alliance, so as to achieve the highest profit of the electricity and gas grids and the overall profit of the alliance.

[0031] The two-layer game model constructed by the present invention guides the flexible adjustment of user-side load through the interaction of upper-layer price information and lower-layer equipment information to achieve the goal of peak shaving and valley filling. At the same time, the alliance reduces energy costs and guarantees the renewable energy absorption capacity in the park to a greater extent.

[0032] (2) The present invention improves the Pathfinder algorithm based on fuzzy control and Archive set methods, solves the operation optimization model of the two-layer game alliance, adaptively selects two control parameters of the Pathfinder algorithm through the fuzzy control method, and combines the external population Archive set method with the Levy flight method to update the position formula. The present invention uses the output of fuzzy control as the key parameter in the Pathfinder algorithm, and adds a population Archive set outside the Pathfinder algorithm to select the optimal individual in each iteration process, which significantly improves the solution efficiency. At the same time, the piecewise linear chaotic mapping method and the improved update position formula are used when initializing the population, which can enhance the randomness and diversity of each individual in the Pathfinder algorithm population, thereby greatly reducing the situation of falling into the local optimum.

[0033] (3) The present invention is based on the Shapley value method, the average strategy allocation method, the nucleolus method and other benefit allocation methods, and uses a weighted method to combine the three methods, evaluates the cooperation willingness of each participating entity through the cumulative prospect method, and weights the benefit allocation method. The weights of the three benefit allocation methods are used as decision variables, and the cumulative prospect theory evaluation is used as the objective function. The optimal weight coefficients of the three allocation methods are solved by the differential evolution algorithm, so that the score obtained by the cumulative prospect theory reaches the optimal value, and the benefit allocation method when the willingness of each participating entity to participate in the alliance is the greatest is obtained, which makes up for the situation that the Shapley value method does not consider unstable factors and causes the alliance to break up, solves the egalitarianism problem generated in the alliance process, and effectively motivates each entity to participate in the collaborative operation of the alliance.

[0034] Other features and advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is an energy flow diagram of the multi-agent cooperative game of the comprehensive energy system in an embodiment of the present invention;

[0036] Figure 2 Schematic diagram of the improved pathfinder algorithm process in an embodiment of the present invention;

[0037] Figure 3 Flow chart of the differential evolution algorithm in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0039] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0040] Embodiment 1

[0041] In one or more embodiments, a multi-agent cooperative game method for an integrated energy system is disclosed, which specifically includes the following process:

[0042] S101: Based on the operating characteristics of the main equipment in the integrated energy system, a combined heat and power park model representing urban suburbs, urban industries, and urban commerce is constructed.

[0043] Specifically, combined Figure 1 The park includes a combined heat and power park, a centralized wind farm, an electric hydrogen production system and an energy storage power station; the electric hydrogen production system purchases electricity from the superior power grid, photovoltaic and wind power generation, and produces hydrogen through electrolyzer equipment. The hydrogen produced first meets the hydrogen load demand in the park. Then part of the hydrogen is used to produce natural gas through methane reactor equipment, and traded to the combined heat and power and combined heat and power parks; the other part produces electric and thermal energy resources through hydrogen fuel cells, and trades them to the combined heat and power, combined heat and power parks and energy storage power stations; there is also electrical energy interaction between the combined heat and power, combined heat and power parks and energy storage power stations, and all three parks use air conditioning to meet the cooling load demand.

[0044] Under the cooperative game operation mode, various stakeholders form an alliance and utilize the characteristics of multi-energy complementarity to achieve the goal of maximizing the benefits of the alliance.

[0045] S102: The upper layer takes the electricity and gas grids as the main body, the comprehensive demand response price as the decision variable, and the maximization of the profits of the electricity and gas grids as the optimization goal; the lower layer takes photovoltaic, wind power, and various parks as the main body, considers the loss costs and carbon quotas in the process of energy transmission between the entities, takes the equipment output of each entity as the decision variable, and takes the maximization of the multi-agent alliance profit as the optimization goal, and establishes an operation optimization model of a two-layer game alliance.

[0046] In this embodiment, considering that the distance between various stakeholders is relatively long, the energy interaction loss between them is shown as formula (1):

[0047] P loss =η e (i,j)D(i,j)P e (i,j)+η h (i,j)D(i,j)P h (i,j) (1)

[0048] Where η e (i,j),η h (i, j) represents the loss rate of electric energy and heat energy transmission of stakeholders i and j respectively; D(i, j) represents the distance between stakeholders i and j; P e (i,j),P h (i, j) represents the power of electrical energy and thermal energy transmission of stakeholders i and j respectively.

[0049] The carbon quotas are as follows:

[0050]

[0051] Where P grid , P gas , Respectively represent the power purchased by each stakeholder from the superior power grid, the amount of natural gas purchased from the superior gas grid, and the overall carbon emissions of the alliance; γ grid , γ gas Represent the carbon emission coefficients of electricity and natural gas respectively.

[0052] The cost of tiered carbon quota trading is shown in formula (3).

[0053]

[0054] In the formula, It represents the cost of step-by-step carbon quota trading; λ represents the base price of carbon quota trading; l represents the length of the step interval of carbon quota trading; β represents the price increase.

[0055] In this embodiment, the operation optimization model of the two-tier game alliance is specifically as follows:

[0056] F upper =F e +F g -F edr -F hdr (4)

[0057]

[0058] In the formula, Fupper 、F lower Represent the objective functions of the upper and lower layers respectively; F e 、F g They represent the income from the upper-level electricity and gas grids selling energy to the lower-level entities; F edr 、F hdr They represent the demand response subsidies given by the upper-level electricity and gas grids to the lower-level parks; F1 represents the income of the distributed photovoltaic park; F2 represents the income of the centralized wind farm; F3 represents the income of the urban industrial cogeneration park; F4 represents the income of the urban commercial cogeneration park; F5 represents the income of the park including the electric hydrogen production system; F6 represents the income of the energy storage power station; F loss represents the loss cost of energy interaction; Represents the transaction cost of tiered carbon quotas.

[0059] At the same time, constraints are established, including upper and lower limits on equipment output, upper limits on transportation capacity of connecting pipelines between parks, and upper limits on energy storage charging and discharging.

[0060] S103: Using the improved Pathfinder algorithm, the operation optimization model of the two-layer game alliance is solved to obtain the maximum multi-agent alliance benefit.

[0061] In this embodiment, the operation optimization model of the double-layer game alliance is solved by improving the Pathfinder algorithm based on fuzzy control and Archive set method to obtain the maximum alliance benefit.

[0062] In the pathfinder algorithm, each member has a position in 2D, 3D or d-dimensional space. At any time, if a member of the group is in the optimal area, it will be selected as the leader. In addition, it is assumed that all candidate solutions to a problem are individual position vectors, and individuals in the group can walk in 2D, 3D and d-dimensional space. In order to find prey or foraging areas, each individual in the population needs to follow the pathfinder.

[0063] In the population update process, the selected leader is the pathfinder of the population's movement direction, moving ahead of other individuals in the population. Its update method is shown in equations (6)-(7):

[0064]

[0065]

[0066] In the formula, iter represents the current number of iterations; iter max Indicates the maximum number of iterations; They respectively represent the last iteration position, current iteration position and updated position information of the pathfinder; R1 and U respectively represent uniformly generated random vectors; s1(A) represents the fluctuation vector, which provides random movement for the pathfinder.

[0067] After the pathfinder completes the position update, other followers update their positions based on the pathfinder. The update method is shown in equations (8)-(11).

[0068]

[0069]

[0070]

[0071] S ij =||s i -s j || (11)

[0072] In the formula, s i and j represents the current position of the follower; R2, R3, V represent uniformly generated random vectors; represents; s2(A) represents the fluctuation vector, providing random movement for the follower; α, β represent the interaction coefficient between followers and the attraction coefficient of the pathfinder to the follower, respectively; S ij represents the distance between the ith follower and the jth follower.

[0073] In the algorithm solution process of this embodiment, the output of fuzzy control is first used as the two key parameter values ​​α and β in the pathfinder algorithm, and the population Archive set is added outside the pathfinder algorithm to select the optimal individual in each iteration process. Then, under the premise of meeting the constraints, the population fitness value is calculated through the equipment output plan, and finally after population iteration, the maximum cooperative alliance benefits are obtained.

[0074] The basic idea of ​​fuzzy control (Mamdani Fuzzy Systems) is to use language variables to represent process states, based on fuzzy set theory, fuzzy language and fuzzy logic, and use these variables as inputs of control rules; the main processes include: input variable fuzzification, fuzzy reasoning and defuzzification.

[0075] During the iteration of the Pathfinder algorithm, the Archive set is used to store the best individuals of each generation. Then after each iteration, the best individuals generated by the new generation will be compared and updated with the best individuals of each previous generation stored in the Archive set.

[0076] If there is no individual in the Archive set that is better than the new individual, then the individual is added to the Archive set; if there is any individual in the Archive set that is better than the new individual, then the individual is not added to the Archive set. At the same time, in order to avoid storing too many similar individuals, the objective function values ​​of all individuals in the Archive set are calculated, and all individuals are arranged according to the distance of the objective function value based on the objective function value, and individuals that are too similar are randomly deleted until the number of individuals in the Archive set reaches the upper limit. Finally, the iterative Archive set is compared with the final optimization result to select the best individual.

[0077] The improved pathfinder algorithm process of this embodiment is shown in the attached figure. Figure 2 As shown, the specific process is as follows:

[0078] S103-1: Taking the comprehensive demand response price of the upper layer as an individual, the algorithm parameters such as the random vectors R1, U, R2, R3 and V of the group, the maximum number of iterations iter are initialized by the piecewise linear chaotic mapping method max , wave vectors s1(A) and s2(A), control parameters α and β, etc.;

[0079] The specific principle of the piecewise linear chaotic mapping method (Piecewise Linear Chaotic Map, PWLCM) is as follows:

[0080]

[0081] In the formula, p m represents the control parameter of the mapping, p m ∈[0,1]; x(t) represents the random iterative value generated.

[0082] S103-2: Initialize the external storage Archive set, and set the number of iterations iter = 1;

[0083] S103-3: Calculate the upper layer objective function for each individual in the initialized population, find the current global optimal solution, set the optimal individual in the initialization and its objective function value as the pathfinder, and store them in the Archive set;

[0084] S103-4: The input variables of the fuzzy system are shown in equations (12)-(13).

[0085]

[0086]

[0087] In the formula, F current , and Respectively represent the objective function values ​​solved by each individual, the best individual, and the worst individual in the Pathfinder algorithm population; F n and They represent the objective function value and average value solved by each individual in the Pathfinder algorithm population respectively; the fuzzification process is to fuzzify the two parameters D and Q, and the values ​​of the control parameters α and β have an impact on the calculation results of D and Q. By fuzzy controlling the two parameters D and Q, the values ​​of the parameters α and β can be obtained;

[0088] S103-5: The fuzzification process uses three pre-defined trapezoidal membership functions, each of which uses three terms: high, medium and low; then combined with the Mamdani fuzzy rules, fuzzy rules for reasoning evaluation are established; the specific membership functions and fuzzy rules can be implemented by technical personnel in this field based on the existing technology, so they are not described in detail.

[0089] S103-6: Use the centroid method to implement the defuzzification process to obtain two control parameters α and β of the Pathfinder algorithm; the defuzzification process can also be implemented by the existing technology.

[0090] S103-7: input the obtained two parameters α and β into the pathfinder algorithm;

[0091] S103-8: According to the improved Levy flight method, the wave vectors s1(A) and s2(A) are calculated as shown in equations (14)-(15).

[0092]

[0093]

[0094] Among them, S lw1 , S lw2 They represent the step lengths obtained by the Levy flight method. This process is implemented by existing technologies and will not be described in detail. iter represents the number of iterations of the pathfinder algorithm. max represents the maximum number of iterations of the Pathfinder algorithm, S ij represents the distance between the i-th follower and the j-th follower.

[0095] S103-9: Calculate the optimal individual of the Pathfinder algorithm population and its objective function value;

[0096] S103-10: Update the location information of each follower;

[0097] S103-11: Compare the new optimal individual with each individual in the Archive set, and determine whether it is stored in the Archive set through the update rule; the specific update rule is: if the objective function value (i.e., profit) corresponding to the new optimal individual is greater than the objective function value corresponding to any individual in the Archive set, then the new optimal individual is retained; if the objective function value (i.e., profit) corresponding to the new optimal individual is less than the minimum value of the objective function values ​​corresponding to all individuals in the Archive set, then it is not retained.

[0098] S103-12: Arrange all individuals according to the distance of the objective function value, and randomly delete individuals that are too close until the number of individuals in the Archive set reaches the upper limit;

[0099] S103-13: Determine whether the number of iterations reaches the upper limit, if yes, execute step S103-15, otherwise execute step S103-14;

[0100] S103-14: number of iterations iter=iter+1, return to step S103-4;

[0101] S103-15: Output the external storage Archive set. For each individual in the Archive set, combine the lower-level objective function to calculate the benefit of the cooperative game alliance, select the individual that maximizes the benefit of the cooperative game alliance, and then determine the device output corresponding to this individual and the benefit of the cooperative game alliance.

[0102] S104: Use the benefit distribution method to distribute the benefits of the multi-agent alliance, and obtain the benefit distribution when the multi-agents have the strongest willingness to participate in the alliance.

[0103] This embodiment is based on the benefit distribution methods such as Shapley value method, average strategy allocation method, and nucleolus method. The weighted method is used to combine the three methods, and the willingness of each participating entity to cooperate is evaluated through the cumulative prospect method to weight the benefit distribution method. The weights of the three benefit distribution methods are used as decision variables, and the cumulative prospect theory evaluation is used as the objective function. The optimal weight coefficients of the three distribution methods are solved by the differential evolution algorithm, so that the score obtained by the cumulative prospect theory reaches the optimal level, and the benefit distribution method when each participating entity is most willing to participate in the alliance is obtained.

[0104] Among them, the allocation idea of ​​the Shapley value method is: the benefit allocation value of each stakeholder in the integrated energy system is equal to the sum of the marginal contributions of the stakeholder to all the alliances in which it participates. The allocation idea of ​​the average strategy allocation method is: each stakeholder participating in the alliance in the integrated energy system will receive the total benefits of the large alliance on average. The idea of ​​the core method allocation is to find a distribution plan that maximizes the satisfaction of all alliance members.

[0105] This embodiment uses the cumulative prospect theory evaluation (maximum prospect value of each solution) as the objective function. The cumulative prospect theory first needs to determine the reference point X ref , X represents the final benefit of each solution, and then determine the benefit value y = XX ref .

[0106] Then determine the value function as shown in formula (16).

[0107]

[0108] In the formula, α and β represent the sensitivity decreasing factors; η represents the loss aversion coefficient; the positive and negative values ​​of Y represent the gains and losses brought about by the decision-making behavior adopted.

[0109] The weight function is determined as shown in equations (17)-(18):

[0110]

[0111]

[0112] Where p represents a weighted function of the probability of an event; μ and ν represent the degree of influence on overestimating low-probability events and underestimating high-probability events, respectively.

[0113] Finally, the prospect value of each solution is calculated as shown in equations (19)-(21):

[0114]

[0115] π + (p i )=w + (p i +…+p n )-w + (p i+1 +…+p n );0≤i≤n-1 (20)

[0116] π - (p i )=w-(p -m +…+p i )-w - (p -m +…+p i-1 );1-m≤i≤0 (21)

[0117] Where CPT(y,p) represents the cumulative prospect value; π + (p i ),π - (p i) represent the positive and negative cumulative decision weight functions, that is, the cumulative decision weight functions when the decision maker faces gains and losses.

[0118] This embodiment uses a differential evolution algorithm to solve the optimal weight coefficients of the three allocation methods. The basic process of the differential evolution algorithm is as follows:

[0119] First, initialize the population and mutation parameters, crossover probability parameters, population number parameters, and individual dimension parameters;

[0120] Then, from the initialized population Randomly select three different individual vectors x i 、x j 、x k , the construction of individual vector is shown in formula (22).

[0121]

[0122] In the formula, D represents the dimension of the individual; N represents the size of the population; iter represents the number of iterations. Represents an individual of the evolutionary algorithm.

[0123] Then perform the mutation operation:

[0124] v=x i +F(x j -x k ) (twenty three)

[0125] Where v represents the calculated mutation vector; F represents the scaling factor.

[0126] Then, a crossover operation is performed to generate diverse offspring vectors, which enhances the diversity of the population, makes the structure more complex, and promotes structural differences in the population. The crossover individual u is calculated through the parent individual x and the mutant individual v, as shown in formula (24).

[0127] u=x+K(vx) (24)

[0128] Where K is the crossover factor, which is used to control the degree of difference between the mutation vector and the parent individual vector.

[0129] Finally, the selection operation is performed by taking the score obtained by the cumulative prospect theory (the prospect value of each plan) as the objective function, directly comparing the parent individual vector x and the mutation vector u, and selecting the better individual as the member of the next generation of individuals, as shown in formula (25).

[0130]

[0131] Repeat the above steps until the optimal weight coefficient is found or the maximum number of iterations is reached.

[0132] Combining the above theoretical knowledge with Figure 3 The specific process of solving the optimal weight coefficients of the three allocation methods by differential evolution algorithm in this embodiment is as follows:

[0133] S S104-1: Initialize differential evolution algorithm parameters such as mutation parameter, crossover probability parameter, population number parameter and individual dimension parameter;

[0134] S S104-2: set the number of iterations iter = 1;

[0135] S S104-3: Use two different vectors in the population to interfere with an existing vector and perform a differential operation to achieve mutation;

[0136] S104-4: Perform crossover to generate test individuals for each individual and its generated offspring mutation vector;

[0137] S104-5: Combine the three benefit distribution methods using the weighted coefficient method, as shown in formula (26) (27):

[0138] x(i)=ax Shapley (i)+bx Average (i)+cx Nucleolus (i) (26)

[0139] a+b+c=1 (27)

[0140] In the formula, x(i) represents the benefits distributed to each stakeholder after participating in the cooperative game alliance; Shapley (i), x Average (i), x Nucleolus (i) represents the benefits obtained by each stakeholder from the alliance according to the Shapley value method, average strategy method and nucleolus method distribution rules; a, b and c are the corresponding weight coefficients.

[0141] S104-6: Assign each weight coefficient as a decision, and calculate the value function under different decisions according to formula (16);

[0142] S104-7: Calculate the weight function according to equations (17) and (18);

[0143] S104-8: Obtain the prospect value under the decision according to equations (19), (20), (21), and select the better one from the target individual and the test individual as the next generation based on the prospect value;

[0144] S104-9: Determine whether the number of iterations reaches the upper limit, if yes, execute step S104-11, otherwise execute step S104-10;

[0145] S104-10: number of iterations iter=iter+1, return to step S104-3;

[0146] S104-11: Select the decision with the largest foreground value as the optimal weight coefficient allocation, thereby outputting the optimal weight coefficient.

[0147] After obtaining the optimal weight coefficient, we can get the benefits distributed to each stakeholder after participating in the cooperative game alliance;

[0148] This embodiment adopts an improved benefit distribution method to reasonably distribute benefits, effectively making up for the situation in which the Shapley value method fails to consider unstable factors and causes the alliance to break up, solves the egalitarianism problem generated in the alliance process, and effectively motivates various entities to participate in the collaborative operation of the alliance.

[0149] Embodiment 2

[0150] In one or more embodiments, a multi-agent cooperative game system for an integrated energy system is disclosed, comprising:

[0151] The integrated energy system model building module is configured to build a combined heat and power park model representing urban suburbs, urban industries, and urban commerce based on the operating characteristics of the main equipment in the integrated energy system;

[0152] The operation optimization model building module is configured as follows: the upper layer takes the electricity and gas grids as the main body, takes the comprehensive demand response price as the decision variable, and takes the maximization of the profit of the electricity and gas grids as the optimization goal; the lower layer takes photovoltaic, wind power, and various parks as the main body, considers the loss cost and carbon quota in the energy transmission process between the main bodies, takes the equipment output of each main body as the decision variable, takes the maximization of the multi-body alliance profit as the optimization goal, and establishes the operation optimization model of the two-layer game alliance;

[0153] The model solving module is configured to use the improved pathfinder algorithm to solve the operation optimization model of the two-layer game alliance to obtain the maximum multi-agent alliance benefit;

[0154] The benefit distribution module is configured to use the benefit distribution method to distribute the benefits of the multi-agent alliance, and obtain the benefit distribution when the multi-agents have the strongest willingness to participate in the alliance.

[0155] The specific implementation of the above modules is the same as that in Example 1 and will not be described in detail.

[0156] Embodiment 3

[0157] In one or more embodiments, a terminal device is disclosed, including a server, the server including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the multi-agent cooperative game method of the integrated energy system in Example 1 when executing the program. For the sake of brevity, it will not be described here.

[0158] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0159] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0160] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software.

[0161] Embodiment 4

[0162] In one or more embodiments, a computer-readable storage medium is disclosed, in which a plurality of instructions are stored, wherein the instructions are suitable for being loaded by a processor of a terminal device and executing the multi-agent cooperative game method for an integrated energy system described in Example 1.

[0163] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A multi-agent cooperative game method for an integrated energy system, characterized in that: include: Based on the operating characteristics of the main equipment in the integrated energy system, a combined heat and power park model representing urban suburbs, urban industry, and urban commerce is constructed; The upper layer takes the electricity and gas grids as the main body, the comprehensive demand response price as the decision variable, and the maximization of the profits of the electricity and gas grids as the optimization goal; The lower layer takes photovoltaic, wind power, and various parks as the main bodies, considers the loss costs and carbon quotas in the energy transmission process between the main bodies, takes the equipment output of each main body as the decision variable, takes the maximum benefit of the multi-main body alliance as the optimization goal, and establishes the operation optimization model of the two-layer game alliance; The improved Pathfinder algorithm is used to solve the operation optimization model of the two-layer game alliance and obtain the maximum multi-agent alliance benefit; The benefits of the multi-agent alliance are distributed using the benefit distribution method, and the benefit distribution is obtained when the multi-agents have the strongest willingness to participate in the alliance.

2. A multi-agent cooperative game method for an integrated energy system as claimed in claim 1, characterized in that: The equipment bodies in the combined heat and power park model include: a combined heat and power park, a centralized wind farm, an electric hydrogen production system and an energy storage power station.

3. A multi-agent cooperative game method for an integrated energy system as claimed in claim 1, characterized in that: The operation optimization model of the two-tier game alliance is specifically as follows: F upper =F e +F g -F edr -F hdr ; Among them, F upper 、F lower Represent the objective functions of the upper and lower layers respectively; F e 、F g They represent the income from the upper-level electricity and gas grids selling energy to the lower-level entities; F edr 、F hdr They represent the demand response subsidies given by the upper-level electricity and gas grids to the lower-level parks; F1 represents the income of the distributed photovoltaic park; F2 represents the income of the centralized wind farm; F3 represents the income of the urban industrial cogeneration park; F4 represents the income of the urban commercial cogeneration park; F5 represents the income of the park including the electric hydrogen production system; F6 represents the income of the energy storage power station; F loss represents the loss cost of energy interaction; Represents the transaction cost of tiered carbon quotas.

4. A multi-agent cooperative game method for an integrated energy system as claimed in claim 1, characterized in that: The upper and lower limits of equipment output, the upper limit of the transportation capacity of the connecting pipelines between parks, and the upper limit of energy storage charging and discharging are used as constraints for the operation optimization model of the two-tier game alliance.

5. A multi-agent cooperative game method for an integrated energy system as claimed in claim 1, characterized in that: The improved Pathfinder algorithm is used to solve the operation optimization model of the two-layer game alliance. The specific process is as follows: Initialize the parameters of the pathfinder algorithm; the initialized parameters include the random vector of the group, the maximum number of iterations, the fluctuation vector and the control parameter; the above-layer comprehensive demand response price is taken as the individual; Calculate the upper layer objective function for each individual in the initialized population, find the current global optimal solution, set the optimal individual in the initialization and its objective function value as the pathfinder, and store them in the Archive set; The two control parameters of the pathfinder algorithm are adaptively selected by the fuzzy control method; the wave vector is calculated according to the Levy flight method; Calculate the optimal individual of the Pathfinder algorithm population and its objective function value, and update the location information of each follower; compare the new optimal individual with each individual in the Archive set to determine whether it is stored in the Archive set, until the number of individuals in the Archive set reaches the upper limit; Output the Archive set, calculate the cooperative game alliance benefits for each individual in the Archive set, and select the individual that maximizes the cooperative game alliance benefits and the corresponding equipment output and optimal benefits.

6. A multi-agent cooperative game method for an integrated energy system as claimed in claim 5, characterized in that: According to the improved calculation of the Levy flight method, the wave vectors s1(A) and s2(A) are as follows: Among them, S lw1 , S lw2 They represent the step size obtained by the Levy flight method, iter represents the number of iterations of the pathfinder algorithm, and iter max represents the maximum number of iterations of the Pathfinder algorithm, S ij Denotes the distance between the ith follower and the jth follower.

7. A multi-agent cooperative game method for an integrated energy system as claimed in claim 1, characterized in that: Use the profit distribution method to distribute the benefits of the multi-agent alliance, specifically: x(i)=ax Shapley (i)+bx Average (i)+cx Nucleolus (i); a+b+c=1; Among them, x(i) represents the benefits distributed to each stakeholder after participating in the cooperative game alliance; x Shapley (i), x Average (i), x Nucleolus (i) represents the benefits obtained by each stakeholder from the alliance according to the Shapley value method, average strategy method and nucleolus method distribution rules; a, b and c are the corresponding weight coefficients; Taking the weight coefficient as the decision variable and the score obtained by cumulative prospect theory as the objective function, the optimal weight coefficients of the three allocation rules are obtained by differential evolution algorithm.

8. A multi-agent cooperative game system for an integrated energy system, characterized in that: include: The integrated energy system model building module is configured to build a combined heat and power park model representing urban suburbs, urban industries, and urban commerce based on the operating characteristics of the main equipment in the integrated energy system; The operation optimization model building module is configured as follows: the upper layer takes the electricity and gas grids as the main body, takes the comprehensive demand response price as the decision variable, and takes the maximization of the profit of the electricity and gas grids as the optimization goal; the lower layer takes photovoltaic, wind power, and various parks as the main body, considers the loss cost and carbon quota in the energy transmission process between the main bodies, takes the equipment output of each main body as the decision variable, takes the maximization of the multi-body alliance profit as the optimization goal, and establishes an operation optimization model of a two-layer game alliance; The model solving module is configured to use the improved pathfinder algorithm to solve the operation optimization model of the two-layer game alliance to obtain the maximum multi-agent alliance benefit; The benefit distribution module is configured to use the benefit distribution method to distribute the benefits of the multi-agent alliance, and obtain the benefit distribution when the multi-agents have the strongest willingness to participate in the alliance.

9. A terminal device, comprising a processor and a memory, wherein the processor is used to implement instructions; and the memory is used to store multiple instructions, characterized in that: The instructions are suitable for being loaded by a processor and executing the multi-agent cooperative game method for an integrated energy system as described in any one of claims 1-7.

10. A computer-readable storage medium storing a plurality of instructions, characterized in that: The instructions are suitable for being loaded by a processor of a terminal device and executing the multi-agent cooperative game method for an integrated energy system as described in any one of claims 1-7.

Images