Inter-provincial electric power spot market declaration strategy optimization method based on quantum genetic algorithm

The quantum genetic algorithm optimizes the inter-provincial power spot market declaration strategy, solves the problems of information asymmetry and poor adaptability of historical data, improves transaction success rate and returns, adapts to market changes on the operation date, and reflects the impact of participants on the spot market.

CN118917461APending Publication Date: 2024-11-08ECONOMIC TECH RES INST OF STATE GRID ANHUI ELECTRIC POWER
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202410949130.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-16
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

In the inter-provincial power spot market, the existing bidding game analysis method has problems such as information asymmetry, historical clearance data cannot adapt to market changes on the day of operation, and ignoring the initiative of the power buyer, resulting in market clearance failure and poor economic performance.

Method used

The inter-provincial power spot market declaration strategy optimization method is adopted based on quantum genetic algorithm. By simulating market dynamic game, the declaration strategy set is generated, the game initiative of the power purchaser is considered, and the strategy diversity is improved by using quantum superposition and full interference crossover to build market equilibrium Point to optimize the declaration strategy.

Benefits of technology

It improves the transaction success rate and returns, adapts to market changes on the day of operation, effectively compensates for information asymmetry, and reflects the impact of participants on the spot market clearance price.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118917461A_ABST
    Figure CN118917461A_ABST
Patent Text Reader

Abstract

The invention relates to an inter-provincial electric power spot market declaration strategy optimization method based on a quantum genetic algorithm. The method comprises the following steps: initializing a group of declaration curves; the simulated market is cleared; constructing a declaration strategy sample generation model of the market participants; calculating a payment matrix of market participants under each combination declaration strategy, initializing an all-zero matrix, solving a market equilibrium point through a static game method, traversing the payment matrix, and marking an index corresponding to a maximum element in the matrix as the market equilibrium point; and carrying out inter-province electric power spot transaction declaration according to the declaration strategy at the market equilibrium point. According to the invention, a better declaration strategy can be obtained in the inter-provincial electric power spot market with incomplete information characteristics, and the transaction success rate and income are improved; the game initiative of the electricity purchaser in the spot market is considered; a strategy sample is simulated in the incomplete information market, a multi-participant game is constructed, and the influence of each participant on the clearing price of the spot market can be better reflected.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of electric power spot market, in particular to an inter-provincial electric power spot market declaration strategy optimization method based on quantum genetic algorithm. Background Art

[0002] In the inter-provincial electricity spot market transactions, the declaration strategy determines the clearing volume and clearing price of the participants in the market. The inter-provincial electricity spot market is an incomplete information market. The bidding strategies declared by the participants in the electricity spot market are mainly generated by sampling probability distributions such as uniform distribution and normal distribution. The blind bidding cannot maximize the benefits for market participants. How to quickly declare a set of bidding strategies that are in line with the interests of the participants based on incomplete information within a limited time has become the main obstacle to the inter-provincial electricity spot market.

[0003] The existing bidding game analysis method for participants in the inter-provincial electricity spot market has the following defects: First, the market is reduced to a two-way game between buyers and sellers. The bilateral game model ignores the differences between individuals, resulting in deviations from the real spot market bidding game scenario; second, the electricity purchaser is regarded as a price acceptor, and only its declared volume is considered in the spot market declaration link, ignoring the game initiative of the electricity purchaser in the market; third, the classic strategies of participants are extracted from historical clearing data, but the spot market has the characteristics of high timeliness, and the historical clearing scenario cannot fully adapt to the market scenario on the operating day, among which the key influences include load fluctuations, unit output fluctuations, changes in grid structure, climate changes, etc. Summary of the invention

[0004] In order to solve the problems that there are deviations between bilateral game scenarios and real spot market game scenarios, the real market clearing results cannot be effectively reflected, and the declaration strategies using historical clearing data as strategy samples for analysis are deviated from the market environment on the operating day, the purpose of the present invention is to provide an inter-provincial electricity spot market declaration strategy optimization method based on a quantum genetic algorithm that can effectively make up for the market clearing failure and poor clearing economic efficiency caused by information asymmetry, and can better reflect the influence of each participant on the spot market clearing price.

[0005] To achieve the above object, the present invention adopts the following technical solution: a method for optimizing the inter-provincial electricity spot market declaration strategy based on quantum genetic algorithm, the method comprising the following steps in order:

[0006] (1) Input a set of market participants’ declared quantities on the operating day [G0, G1, …, G n ], where G0 represents the total amount of daily declarations of market participants numbered 0, and initializes a set of declaration curves;

[0007] (2) Simulate market clearing, use the initialized set of declaration curves as the initial strategy of each market participant, perform orderly clearing according to the marginal clearing rule, and feed back the clearing results to each market participant;

[0008] (3) Based on the clearing results, a model for generating sample declaration strategies of market participants is constructed, and the declaration strategies for the next round are optimized through the quantum genetic algorithm. The dynamic market game is simulated to generate a declaration strategy set. In the dynamic game process, the declaration strategies of the market participants in the previous round are used as the input of the quantum genetic algorithm.

[0009] (4) Using clustering methods to reduce the dimension of the application strategy set, extracting multiple classic application strategies from the application strategy set to form a classic application strategy set;

[0010] (5) Select a classic declaration strategy from each market participant's classic declaration strategy set to form a combination declaration strategy, and calculate the payoff matrix of the market participants under each combination declaration strategy. The form of the payoff matrix is ​​a0×a1×a2×…a n ×b, where a0, a1, a2, a n is the number of classic declaration strategies of n+1 market participants; b is the payoff matrix, i.e. [b1, b2, …, b n+1 ], used to store the returns of each market participant under the portfolio declaration strategy;

[0011] (6) Initialize an all-zero matrix as the marker matrix for finding the market equilibrium point. The format of the marker matrix is ​​a0×a1×a2×…a n , where a0, a1, a2…a n is the number of classic declared strategies of n+1 market participants;

[0012] (7) Solve the market equilibrium point through the static game method, traverse the payment matrix, and record b1, b2 to b in the payoff matrix. n+1 The index of the payment matrix corresponding to the maximum value (s0, s1, ...s n+1 ), and the index in the marker matrix is ​​(s0, s1, ...s n+1 ) plus 1, and the index corresponding to the largest element in the marker matrix is ​​the market equilibrium point;

[0013] (8) Conduct inter-provincial electricity spot trading declarations according to the declaration strategy at the market equilibrium point.

[0014] In step (1), the set of declared curves is:

[0015] [s i1 , j i1 ,s i2 , j i2 ,si3 , j i3 ,s i4 , j i4 ,s i5 , j i5 ]

[0016] Among them, s i1 、s i2 、s i3 、s i4 、s i5 The market participant numbered i declares the amount of each segment of the curve in segments, satisfying s i1 +s i2 +s i3 +s i4 +s i5 =G i , j i1 、j i2 、j i3 、j i4 、j i5 The s in the segmented declaration curve are i1 、s i2 、s i3 、s i4 、s i5 The corresponding declared price; i is the market participant number, i∈Ω, Ω is the set of market participants; G i Represents the total amount reported on the operating day by market participant numbered i.

[0017] In step (2), the clearing result includes the market marginal clearing price price i,clear 、The current round of market clearing electricity P i,clear , as well as the current round declaration curve of each market participant.

[0018] In step (3), the objective function of the declaration strategy sample generation model includes an optimization objective function of the electricity seller and an optimization objective function of the electricity buyer. The optimization objective function of the electricity seller is:

[0019] max[P i,clear ×max(price 0,i,t )]

[0020] Where P i,clear The electricity volume cleared in this round of market, price 0,i,t is the marginal clearing price of this round of market, max[P i,clear ×max(price 0,i,t )] indicates that the optimization goal of the electricity seller is to clear the most electricity when the marginal clearing price of this round of market is the largest;

[0021] The optimization objective function of the electricity purchaser is:

[0022] max[P i,clear ×min(price 0,i,t )]

[0023] In the formula, max[P i,clear ×min(price 0,i,t )] indicates that the optimization goal of the power purchaser is to clear the most electricity in this round when the marginal clearing price of the market is the lowest, with P i,clear ×price 0,i,t As a market participant, adaptability;

[0024] The constraints of the application strategy sample generation model include line flow constraints, upper and lower limit constraints of power sales party generation, and upper and lower limit constraints of power purchase party load, wherein the line flow constraints are:

[0025] -l v,k ≤F v,k ≤l v,k

[0026] The upper and lower limits of the electricity sales party’s power generation are:

[0027] P min,i ≤P i ≤P max,i

[0028] The upper and lower limits of the power purchaser's load are:

[0029] L min,j ≤L j ≤L max,j

[0030] In the formula, l v,k is the transmission capacity of the kth inter-provincial connection line, P min,i and P max,i are the minimum and maximum output of the ith electricity seller, respectively, L min,j and L max,j are the minimum load and maximum load of the jth power purchaser respectively; F v,k is the power flow of the kth inter-provincial connection line on the operation day, P i is the electricity quantity reported by the electricity seller on operation day i, L j The electricity quantity reported by the electricity purchaser on operation day numbered j.

[0031] In step (3), constructing a declaration strategy sample generation model specifically includes the following steps in order:

[0032] (3a) Initialize the basic parameters of the quantum genetic algorithm, including the population size Δ, the number of quantum registers d, the maximum number of iterations τ, the length of the quantum chromosome γ, the number of quantum register qubits c, the quantum rotation angle ε, and the number of quantum chromosomes and the number of optimal solutions v; d is an integer multiple of 4, c is an even number; in the context of segmented application, the following relationship is satisfied:

[0033]

[0034] (3b) Initialize Δ quantum registers, each of which represents a candidate solution and contains c quantum bits;

[0035] (3c) Initialize Δ quantum registers to random quantum states, representing the initial population:

[0036]

[0037] Where i is the market participant number, i∈Ω, Ω is the set of market participants; δ i is the gene encoded in the ith quantum register, c il For gene δ i The lth quantum bit state, l represents the number of quantum bits used by a quantum gene;

[0038] (3d) Measure the state of the quantum register and obtain the binary gene [c i1 c i2 …c il ], convert the binary gene into a decimal number and get the declaration strategy:

[0039]

[0040] Among them, P i,max , P i,min They represent the highest and lowest declared amount in each period of market participant i, J i,max , J i,min represents the highest and lowest bid prices of each segment of market participant i, p represents the binary gene number; S ip,m is the first Segment declaration quantity, c ip For binary genes [c i1 c i2 ...c il ], j ip,e is the first The declared price corresponding to the declared quantity of the segment;

[0041] Through market clearing, we get the market marginal clearing price under the current declaration strategy. i,clear 、The current round of market clearing electricity P i,clear , calculate the current expenditure under clearing, and traverse all individuals in the population;

[0042] (3e) sorting the plurality of quantum registers in descending order of the fitness values ​​of the market participants, wherein the fitness values ​​of the market participants of the first half of the quantum registers are higher than the fitness values ​​of the market participants of the second half of the quantum registers, resetting the second half of the quantum registers to the reference state ρ0, and using a quantum cloning machine QCM to copy the quantum bits of the first half of the quantum registers to the second half of the quantum registers;

[0043] (3f) Perform crossover operation in the second half of the quantum register, specifically: traverse all a and use quantum exchange gates to After the quantum register qubit, and After the quantum register The quantum bits are exchanged.

[0044]

[0045] (3g) with the same probability p m Perform quantum bit rotation on each qubit in the second half of the quantum register to achieve quantum gene mutation. Specifically, use R x , R y , R z Gates operate on qubits in a quantum register:

[0046]

[0047]

[0048]

[0049] In the formula, θ represents the rotation angle of the quantum bit;

[0050] (3h) Repeat steps (3d) to (3g) until the maximum number of iterations τ is reached;

[0051] (3i) For the declaration strategies obtained by encoding each generation of individuals in the iterative process, the objective function of the declaration strategy sample generation model is used to calculate the market participant's benefits, and the optimal declaration strategy is selected. The market participant's fitness value under the optimal declaration strategy is the highest, and the optimal declaration strategy is retained as the market participant declaration sample;

[0052] (3j) Simulate the dynamic market game process and set the sample generation number P max , when the algorithm iterates to P max When it ends, the declaration strategy set is output, and the declaration strategy set is n×P max A 1×10 array, where n is the number of market participants.

[0053] The step (4) specifically comprises the following steps:

[0054] (4a) For market participant i, i∈Ω, Ω is the set of market participants, calculate the hn+ith, h∈{N|[hn+i]≤P max The average silhouette coefficient of each declared strategy under different numbers of cluster centers, and the number of cluster centers t corresponding to the average silhouette coefficient; P max is the sample generation number, n is the number of market participants, and N is a set of natural numbers;

[0055] (4b) Use K-Means clustering method to cluster the hn+ith ... max} declaration strategies are clustered, with t as the number of cluster centers. The t cluster centers output by the K-Means clustering method constitute the classic declaration strategy set of market participant i.

[0056] The step (7) specifically comprises the following steps:

[0057] (7a) When the first market participant selects the first classical strategy from its classical strategy set, it finds the maximum value of b1 in the payoff matrix b and records the current payoff matrix index as:

[0058] (s0, s1, ...s n+1 )

[0059] Where s0∈Z ∩[1,a0], s1∈Z ∩[1,a1], …, s n+1 ∈Z ∩[1, a n+1 ]; Z is a set of positive integers;

[0060] And let the index in the marker matrix be (s0, s1, ...s n+1 ) plus 1 for the corresponding element;

[0061] (7b) When the first market participant selects the first classical strategy from its classical strategy set, find the maximum value of b2 in the payoff matrix b, record the current payoff matrix index, and add 1 to the element in the flag matrix corresponding to the index. In the same way, find b3, b4, and so on in the payoff matrix b when the first market participant selects the first classical strategy. n+1 The maximum value is obtained by recording the payment matrix index and adding 1 to the position element corresponding to the index in the flag matrix.

[0062] (7c) traverse the first market participant in the same way as step (7a) and step (7b) to select the maximum value of all elements in the payoff matrix b under all its classical strategies, and record the index, so that the element corresponding to the index position in the mark matrix is ​​increased by 1;

[0063] (7d) Find the index of the element with the largest value in the marker matrix (s σ ,s 1σ1 ,…,s nσn ), at the same time, [s σ ,s 1σ1 ,…,s nσn ] is the strategy combination at market equilibrium, which means that the first market participant chooses the σth classic strategy to declare, the second market participant chooses the σ1th classic strategy to declare, and the nth market participant chooses the σth classic strategy to declare. n A classic strategy is declared and the market is balanced.

[0064] It can be seen from the above technical scheme that the beneficial effects of the present invention are as follows: first, the present invention constructs a quantum genetic algorithm encoded in an independent quantum register, and utilizes the characteristics of quantum superposition so that each gene can contain more information, greatly enriching the diversity of the population. On the other hand, the gene mutation method of quantum full interference crossover is used to further improve the probability of population evolution, which can avoid the disadvantage that traditional genetic algorithms are prone to fall into local optimality, and can derive a better declaration strategy in the inter-provincial power spot market with incomplete information characteristics, thereby improving the transaction success rate and benefits; second, compared with the same type of technology, the present invention can obtain a more in line with the competition situation on the operation day. The declaration strategy under the environment, the existing technology obtains the classic strategy through historical transaction data, which cannot be better applied to the operation day declaration plan. The reasons include changes in the grid structure, climate changes, and the increase in the number of participants. The present invention takes into account the game initiative of the power purchaser in the spot market. Thirdly, the present invention simulates the strategy samples in the incomplete information market, which can effectively make up for the market clearing failure and poor clearing economy caused by information asymmetry. On the other hand, the present invention constructs a multi-participant game, and each participant declares a quantity-price combination, which can better reflect the impact of each participant on the spot market clearing price. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 is a flow chart of the method of the present invention;

[0066] Figure 2 It is a flow chart of the method for generating declaration strategy samples based on quantum genetic algorithm of the present invention. DETAILED DESCRIPTION

[0067] like Figure 1 As shown, a method for optimizing the declaration strategy of the inter-provincial electricity spot market based on a quantum genetic algorithm comprises the following steps in order:

[0068] (1) Input a set of market participants’ declared quantities on the operating day [G0, G1, …, G n], where G0 represents the total amount of daily declarations of market participants numbered 0, and initializes a set of declaration curves;

[0069] (2) Simulate market clearing, use the initialized set of declaration curves as the initial strategy of each market participant, perform orderly clearing according to the marginal clearing rule, and feed back the clearing results to each market participant;

[0070] (3) Based on the clearing results, a model for generating sample declaration strategies of market participants is constructed, and the declaration strategies for the next round are optimized through the quantum genetic algorithm. The dynamic market game is simulated to generate a declaration strategy set. In the dynamic game process, the declaration strategies of the market participants in the previous round are used as the input of the quantum genetic algorithm.

[0071] (4) Using clustering methods to reduce the dimension of the application strategy set, extracting multiple classic application strategies from the application strategy set to form a classic application strategy set;

[0072] (5) Select a classic declaration strategy from each market participant's classic declaration strategy set to form a combination declaration strategy, and calculate the payoff matrix of the market participants under each combination declaration strategy. The form of the payoff matrix is ​​a0×a1×a2×…a n ×b, where a0, a1, a2, a n is the number of classic declaration strategies of n+1 market participants; b is the payoff matrix, i.e. [b1, b2, …, b n+1 ], used to store the returns of each market participant under the portfolio declaration strategy;

[0073] (6) Initialize an all-zero matrix as the marker matrix for finding the market equilibrium point. The format of the marker matrix is ​​a0×a1×a2×…a n , where a0, a1, a2…a n is the number of classic declared strategies of n+1 market participants;

[0074] (7) Solve the market equilibrium point through the static game method, traverse the payment matrix, and record b1, b2 to b in the payoff matrix. n+1 The index of the payment matrix corresponding to the maximum value (s0, s1, ...s n+1 ), and the index in the marker matrix is ​​(s0, s1, ...s n+1 ) plus 1, and the index corresponding to the largest element in the marker matrix is ​​the market equilibrium point;

[0075] (8) Conduct inter-provincial electricity spot trading declarations according to the declaration strategy at the market equilibrium point.

[0076] In step (1), the set of declared curves is:

[0077] [si1 , j i1 ,s i2 , j i2 ,s i3 , j i3 ,s i4 , j i4 ,s i5 , j i5 ]

[0078] Among them, s i1 、s i2 、s i3 、s i4 、s i5 The market participant numbered i declares the amount of each segment of the curve in segments, satisfying s i1 +s i2 +s i3 +s i4 +s i5 =G i , j i1 、j i2 、j i3 、j i4 、j i5 The s in the segmented declaration curve are i1 、s i2 、s i3 、s i4 、s i5 The corresponding declared price, where the electricity seller follows the price monotonically non-decreasing curve, and the electricity buyer follows the price monotonically non-increasing curve. i is the market participant number, i∈Ω, Ω is the set of market participants; G i Represents the total amount reported on the operating day by market participant numbered i.

[0079] In step (2), the clearing result includes the market marginal clearing price price i,clear 、The current round of market clearing electricity P i,clear , as well as the current round declaration curve of each market participant.

[0080] In step (3), the objective function of the declaration strategy sample generation model includes an optimization objective function of the electricity seller and an optimization objective function of the electricity buyer. The optimization objective function of the electricity seller is:

[0081] max[P i,clear ×max(price 0,i,t )]

[0082] Where P i,clear The electricity volume cleared in this round of market, price 0,i,t is the marginal clearing price of this round of market, max[P i,clear ×max(price0,i,t )] indicates that the optimization goal of the electricity seller is to clear the most electricity when the marginal clearing price of this round of market is the largest;

[0083] The optimization objective function of the electricity purchaser is:

[0084] max[P i,clear ×min(price 0,i,t )]

[0085] In the formula, max[P i,clear ×min(price 0,i,t )] indicates that the optimization goal of the power purchaser is to clear the most electricity in this round when the marginal clearing price of the market is the lowest, with P i,clear ×price 0,i,t As a market participant, adaptability;

[0086] The constraints of the application strategy sample generation model include line flow constraints, upper and lower limit constraints of power sales party generation, and upper and lower limit constraints of power purchase party load, wherein the line flow constraints are:

[0087] -l v,k ≤F v,k ≤l v,k

[0088] The upper and lower limits of the electricity sales party’s power generation are:

[0089] P min,i ≤P i ≤P max,i

[0090] The upper and lower limits of the power purchaser's load are:

[0091] L min,j ≤L j ≤L max,j

[0092] In the formula, l v,k is the transmission capacity of the kth inter-provincial connection line, P min,i and P max,i are the minimum and maximum output of the ith electricity seller, respectively, L min,j and L max,j are the minimum load and maximum load of the jth power purchaser respectively; F v,k is the power flow of the kth inter-provincial connection line on the operation day, P i is the electricity quantity reported by the electricity seller on operation day i, L j The electricity quantity reported by the electricity purchaser on operation day numbered j.

[0093] In step (3), if Figure 2 As shown in the figure, constructing a declaration strategy sample generation model specifically includes the following steps in order:

[0094] (3a) Initialize the basic parameters of the quantum genetic algorithm, including the population size Δ, the number of quantum registers d, the maximum number of iterations τ, the length of the quantum chromosome γ, the number of quantum register qubits c, the quantum rotation angle ε, and the number of quantum chromosomes and the number of optimal solutions v; d is an integer multiple of 4, c is an even number; in the context of segmented application, the following relationship is satisfied:

[0095]

[0096] (3b) Initialize Δ quantum registers, each of which represents a candidate solution and contains c quantum bits;

[0097] (3c) Initialize Δ quantum registers to random quantum states, representing the initial population:

[0098]

[0099] Where i is the market participant number, i∈Ω, Ω is the set of market participants; δ i is the gene encoded in the ith quantum register, c il For gene δ i The lth quantum bit state, l represents the number of quantum bits used by a quantum gene;

[0100] (3d) Measure the state of the quantum register and obtain the binary gene [c i1 c i2 …c il ], convert the binary gene into a decimal number and get the declaration strategy:

[0101]

[0102] Among them, P i,max , P i,min They represent the highest and lowest declared amount in each period of market participant i, J i,max , J i,min represents the highest and lowest bid prices of each segment of market participant i, p represents the binary gene number; S ip,m is the first Segment declaration quantity, c ip For binary genes [c i1 c i2 ...c il ], j ip,e is the first The declared price corresponding to the declared quantity of the segment;

[0103] Through market clearing, we get the market marginal clearing price under the current declaration strategy. i,clear 、The current round of market clearing electricity P i,clear , calculate the current expenditure under clearing, and traverse all individuals in the population;

[0104] (3e) sorting the plurality of quantum registers in descending order of the fitness values ​​of the market participants, wherein the fitness values ​​of the market participants of the first half of the quantum registers are higher than the fitness values ​​of the market participants of the second half of the quantum registers, resetting the second half of the quantum registers to the reference state ρ0, and using a quantum cloning machine QCM to copy the quantum bits of the first half of the quantum registers to the second half of the quantum registers;

[0105] (3f) Perform crossover operation in the second half of the quantum register, specifically: traverse all a and use quantum exchange gates to After the quantum register qubit, and After the quantum register The quantum bits are exchanged.

[0106]

[0107] (3g) with the same probability p m Perform quantum bit rotation on each qubit in the second half of the quantum register to achieve quantum gene mutation. Specifically, use R x , R y , R z Gates operate on qubits in a quantum register:

[0108]

[0109]

[0110]

[0111] In the formula, θ represents the rotation angle of the quantum bit;

[0112] (3h) Repeat steps (3d) to (3g) until the maximum number of iterations τ is reached;

[0113] (3i) For the declaration strategies obtained by encoding each generation of individuals in the iterative process, the objective function of the declaration strategy sample generation model is used to calculate the market participant's benefits, and the optimal declaration strategy is selected. The market participant's fitness value under the optimal declaration strategy is the highest, and the optimal declaration strategy is retained as the market participant declaration sample;

[0114] (3j) Simulate the dynamic market game process and set the sample generation number P max , when the algorithm iterates to P max When it ends, the declaration strategy set is output, and the declaration strategy set is n×P max A 1×10 array, where n is the number of market participants.

[0115] The step (4) specifically comprises the following steps:

[0116] (4a) For market participant i, i∈Ω, Ω is the set of market participants, calculate the hn+ith, h∈{N|[hn+i]≤P max The average silhouette coefficient of each declared strategy under different numbers of cluster centers, and the number of cluster centers t corresponding to the average silhouette coefficient; P max is the sample generation number, n is the number of market participants, and N is a set of natural numbers;

[0117] (4b) Use K-Means clustering method to cluster the hn+ith ... max} declaration strategies are clustered, with t as the number of cluster centers. The t cluster centers output by the K-Means clustering method constitute the classic declaration strategy set of market participant i.

[0118] The step (7) specifically comprises the following steps:

[0119] (7a) When the first market participant selects the first classical strategy from its classical strategy set, it finds the maximum value of b1 in the payoff matrix b and records the current payoff matrix index as:

[0120] (s0, s1, ...s n+1 )

[0121] Where s0∈Z∩[1,a0], s1∈Z∩[1,a1], …, s n+1 ∈Z∩[1,a n+1 ]; Z is a set of positive integers;

[0122] And let the index in the marker matrix be (s0, s1, ...s n+1 ) plus 1 for the corresponding element;

[0123] (7b) When the first market participant selects the first classical strategy from its classical strategy set, find the maximum value of b2 in the payoff matrix b, record the current payoff matrix index, and add 1 to the element in the flag matrix corresponding to the index. In the same way, find b3, b4, and so on in the payoff matrix b when the first market participant selects the first classical strategy. n+1The maximum value is obtained by recording the payment matrix index and adding 1 to the position element corresponding to the index in the flag matrix.

[0124] (7c) traverse the first market participant in the same way as step (7a) and step (7b) to select the maximum value of all elements in the payoff matrix b under all its classical strategies, and record the index, so that the element corresponding to the index position in the mark matrix is ​​increased by 1;

[0125] (7d) Find the index of the element with the largest value in the marker matrix (s σ ,s 1σ1 ,…,s nσn ), at the same time, [s σ ,s 1σ1 ,…,s nσn ] is the strategy combination at market equilibrium, which means that the first market participant chooses the σth classic strategy to declare, the second market participant chooses the σ1th classic strategy to declare, and the nth market participant chooses the σth classic strategy to declare. n A classic strategy is declared and the market is balanced.

[0126] In summary, the present invention constructs a quantum genetic algorithm encoded in an independent quantum register, and utilizes the characteristics of quantum superposition so that each gene can contain more information, greatly enriching the diversity of the population. On the other hand, the gene mutation method of quantum full interference crossover is used to further improve the probability of population evolution, which can avoid the disadvantage that traditional genetic algorithms are prone to fall into local optimality, and can obtain a better declaration strategy in the inter-provincial power spot market with incomplete information characteristics, thereby improving the transaction success rate and benefits. Compared with similar technologies, the present invention can obtain a declaration strategy that is more in line with the competition scenario of the operation day. The existing technology obtains classical strategies through historical transaction data, which cannot be better applied to the operation day declaration plan. The reasons include changes in grid structure, climate changes, and an increase in the number of participants. The present invention takes into account the game initiative of the power purchaser in the spot market. The present invention simulates a strategy sample in an incomplete information market, which can effectively make up for the market clearing failure and poor clearing economy caused by information asymmetry. On the other hand, the present invention constructs a multi-participant game, and each participant declares a quantity-price combination, which can better reflect the influence of each participant on the spot market clearing price.

Claims

1. A method for optimizing inter-provincial electricity spot market declaration strategy based on quantum genetic algorithm, characterized by: The method comprises the following steps in order: (1) Input a set of market participants’ declared quantities on the operating day [G0, G1, …, G n ], where G0 represents the total amount of daily declarations of market participants numbered 0, and initializes a set of declaration curves; (2) Simulate market clearing, use the initialized set of declaration curves as the initial strategy of each market participant, perform orderly clearing according to the marginal clearing rule, and feed back the clearing results to each market participant; (3) Based on the clearing results, a model for generating sample declaration strategies of market participants is constructed, and the declaration strategies for the next round are optimized through the quantum genetic algorithm. The dynamic game of the market is simulated to generate a declaration strategy set. In the dynamic game process, the declaration strategies of the market participants in the previous round are used as the input of the quantum genetic algorithm. (4) Using clustering methods to reduce the dimension of the application strategy set, extracting multiple classic application strategies from the application strategy set to form a classic application strategy set; (5) Select a classic declaration strategy from each market participant's classic declaration strategy set to form a combination declaration strategy, and calculate the payoff matrix of the market participants under each combination declaration strategy. The form of the payoff matrix is ​​a0×a1×a2×…a n ×b, where a0, a1, a2, a n is the number of classic declaration strategies of n+1 market participants; b is the payoff matrix, i.e. [b1,b2,…,b n+1 ], used to store the returns of each market participant under the portfolio declaration strategy; (6) Initialize an all-zero matrix as the marker matrix for finding the market equilibrium point. The format of the marker matrix is ​​a0×a1×a2×…a n , where a0, a1, a2…a n is the number of classic declared strategies of n+1 market participants; (7) Solve the market equilibrium point through the static game method, traverse the payment matrix, and record b1, b2 to b in the payoff matrix n+1 The index of the payment matrix corresponding to the maximum value (s0, s1, ...s n+1 ), and the index in the marker matrix is ​​(s0,s1,…s n+1 ) plus 1, and the index corresponding to the largest element in the marker matrix is ​​the market equilibrium point; (8) Conduct inter-provincial electricity spot trading declarations according to the declaration strategy at the market equilibrium point.

2. The inter-provincial electricity spot market declaration strategy optimization method based on quantum genetic algorithm according to claim 1 is characterized by: In step (1), the set of declared curves is: [s i1 ,j i1 ,s i2 ,j i2 ,s i3 ,j i3 ,s i4 ,j i4 ,s i5 ,j i5 ] Among them, s i1 、s i2 、s i3 、s i4 、s i5 The market participant numbered i declares the amount of each segment of the curve in segments, satisfying s i1 +s i2 +s i3 +s i4 +s i5 =G i , j i1 、j i2 、j i3 、j i4 、j i5 The s in the segmented declaration curve are i1 、s i2 、s i3 、s i4 、s i5 The corresponding declared price; i is the market participant number, i∈Ω, Ω is the set of market participants; G i Represents the total amount reported on the operating day by market participant numbered i.

3. The inter-provincial electricity spot market declaration strategy optimization method based on quantum genetic algorithm according to claim 1 is characterized by: In step (2), the clearing result includes the market marginal clearing price price i,clear 、The current round of market clearing electricity P i,clear , as well as the current round declaration curve of each market participant.

4. The inter-provincial electricity spot market declaration strategy optimization method based on quantum genetic algorithm according to claim 1 is characterized by: In step (3), the objective function of the declaration strategy sample generation model includes an optimization objective function of the electricity seller and an optimization objective function of the electricity buyer. The optimization objective function of the electricity seller is: max[P i,clear ×max(price 0,i,t )] Where P i,clear The electricity volume cleared in this round of market, price 0,i,t is the marginal clearing price of the current market, max[P i,clear ×max(price 0,i,t )] indicates that the optimization goal of the electricity seller is to clear the most electricity when the marginal clearing price of this round of market is the largest; The optimization objective function of the electricity purchaser is: max[P i,clear ×min(price 0,i,t )] In the formula, max[P i,clear ×min(price 0,i,t )] indicates that the optimization goal of the power purchaser is to clear the most electricity in this round when the marginal clearing price of the market is the lowest, with P i,clear ×price 0,i,t As a market participant, adaptability; The constraints of the application strategy sample generation model include line flow constraints, upper and lower limit constraints of power sales party generation, and upper and lower limit constraints of power purchase party load, wherein the line flow constraints are: -l v,k ≤F v,k ≤l v,k The upper and lower limits of the electricity sales party’s power generation are: P min,i ≤P i ≤P max,i The upper and lower limits of the power purchaser's load are: L min,j ≤L j ≤L max,j In the formula, l v,k is the transmission capacity of the kth inter-provincial connection line, P min,i and P max,i are the minimum and maximum output of the ith electricity seller, respectively, L min,j and L max,j are the minimum load and maximum load of the jth power purchaser respectively; F v,k is the power flow of the kth inter-provincial connection line on the operation day, P i is the electricity quantity reported by the electricity seller on operation day i, L j The electricity quantity reported by the electricity purchaser on operation day numbered j.

5. The inter-provincial electricity spot market declaration strategy optimization method based on quantum genetic algorithm according to claim 1 is characterized by: In step (3), constructing a declaration strategy sample generation model specifically includes the following steps in order: (3a) Initialize the basic parameters of the quantum genetic algorithm, including the population size Δ, the number of quantum registers d, the maximum number of iterations τ, the length of the quantum chromosome γ, the number of quantum register qubits c, the quantum rotation angle ε, and the number of quantum chromosomes and the number of optimal solutions v; d is an integer multiple of 4, c is an even number; in the context of segmented application, the following relationship is satisfied: (3b) Initialize Δ quantum registers, each of which represents a candidate solution and contains c quantum bits; (3c) Initialize Δ quantum registers to random quantum states, representing the initial population: Where i is the market participant number, i∈Ω, Ω is the set of market participants; δ i is the gene encoded in the ith quantum register, c il For gene δ i The lth quantum bit state, l represents the number of quantum bits used by a quantum gene; (3d) Measure the state of the quantum register and obtain the binary gene [c i1 c i2 …c il ], convert the binary gene into a decimal number and get the declaration strategy: Among them, P i,max , P i,min They represent the highest and lowest declared amount in each period of market participant i, J i,max , J i,min represents the highest and lowest bid prices of each segment of market participant i, p represents the binary gene number; S ip,m is the first Segment declaration quantity, c ip For binary genes [c i1 c i2 …c il ], j ip,e is the first The declared price corresponding to the declared quantity of the segment; Through market clearing, we get the market marginal clearing price under the current declaration strategy. i,clear 、The current round of market clearing electricity P i,clear , calculate the current expenditure under clearing, and traverse all individuals in the population; (3e) sorting the plurality of quantum registers in descending order of the fitness values ​​of the market participants, wherein the fitness values ​​of the market participants of the first half of the quantum registers are higher than the fitness values ​​of the market participants of the second half of the quantum registers, resetting the second half of the quantum registers to the reference state ρ0, and using a quantum cloning machine QCM to copy the quantum bits of the first half of the quantum registers to the second half of the quantum registers; (3f) Perform crossover operation in the second half of the quantum register, specifically: traverse all a and use quantum exchange gates to After the quantum register qubit, and After the quantum register The quantum bits are exchanged. (3g) with the same probability p m Perform quantum bit rotation on each qubit in the second half of the quantum register to achieve quantum gene mutation. Specifically, use R x , R y , R z Gates operate on qubits in a quantum register: In the formula, θ represents the rotation angle of the quantum bit; (3h) Repeat steps (3d) to (3g) until the maximum number of iterations τ is reached; (3i) For the declaration strategies obtained by encoding each generation of individuals in the iterative process, the objective function of the declaration strategy sample generation model is used to calculate the market participant's benefits, and the optimal declaration strategy is selected. The market participant's fitness value under the optimal declaration strategy is the highest, and the optimal declaration strategy is retained as the market participant declaration sample; (3j) Simulate the dynamic market game process and set the sample generation number P max , when the algorithm iterates to P max When it ends, the declaration strategy set is output, and the declaration strategy set is n×P max A 1×10 array, where n is the number of market participants.

6. The inter-provincial electricity spot market declaration strategy optimization method based on quantum genetic algorithm according to claim 1 is characterized by: The step (4) specifically comprises the following steps: (4a) For market participant i, i∈Ω, Ω is the set of market participants, calculate the first hn+i,h∈{N|[hn+i]≤P in the declared strategy set max The average silhouette coefficient of each declared strategy under different numbers of cluster centers, and the number of cluster centers t corresponding to the average silhouette coefficient; P max is the sample generation number, n is the number of market participants, and N is a set of natural numbers; (4b) Use K-Means clustering method to cluster the hn+i,h∈{N|[hn+i]≤P of the declared strategy set of market participant i. max } declaration strategies are clustered, with t as the number of cluster centers. The t cluster centers output by the K-Means clustering method constitute the classic declaration strategy set of market participant i.

7. The inter-provincial electricity spot market declaration strategy optimization method based on quantum genetic algorithm according to claim 1 is characterized by: The step (7) specifically comprises the following steps: (7a) When the first market participant selects the first classical strategy from its classical strategy set, it finds the maximum value of b1 in the payoff matrix b and records the current payoff matrix index as: (s0,s1,…s n+1 ) Where s0∈Z∩[1,a0],s1∈Z∩[1,a1],…,s n+1 ∈Z∩[1,a n+1 ]; Z is a set of positive integers; And let the index in the marker matrix be (s0,s1,…s n+1 ) plus 1 for the corresponding element; (7b) When the first market participant selects the first classical strategy from its classical strategy set, find the maximum value of b2 in the payoff matrix b, record the current payoff matrix index, and add 1 to the element in the flag matrix corresponding to the index. In the same way, find b3, b4, and so on in the payoff matrix b when the first market participant selects the first classical strategy. n+1 The maximum value is obtained by recording the payment matrix index and adding 1 to the position element corresponding to the index in the flag matrix. (7c) traverse the first market participant in the same way as step (7a) and step (7b) to select the maximum value of all elements in the payoff matrix b under all its classical strategies, and record the index, so that the element corresponding to the index position in the mark matrix is ​​increased by 1; (7d) Find the index of the element with the largest value in the marker matrix (s σ ,s 1σ1 ,…,s nσn ), at the same time, [s σ ,s 1σ1 ,…,s nσn ] is the strategy combination at market equilibrium, which means that the first market participant chooses the σth classic strategy to declare, the second market participant chooses the σ1th classic strategy to declare, and the nth market participant chooses the σth classic strategy to declare. n A classic strategy is declared and the market is balanced.

Citation Information

Cited By

  • Physical transaction quantum optimization modeling method and system of transformer area cluster, electronic equipment and medium

    CN122048519A

  • Physical transaction quantum optimization modeling method, system, electronic equipment and medium for transformer area cluster

    CN122048519B