Cascade reservoir group flood control scheduling optimization method based on improved flood optimization algorithm

Abstract

The invention discloses a cascade reservoir group flood control scheduling optimization method based on an improved flood optimization algorithm. The method comprises the steps of obtaining cascade reservoir group flood control scheduling basic data; determining an initial calculation condition, and establishing a fitness function; performing population initialization operation to obtain a population comprising a plurality of individuals and obtain an optimal individual of the population; population optimization operation: updating the population and the optimal population individual through flood routing simulation operation, flood change simulation operation, elite reverse learning operation and longitudinal cross learning operation in sequence; and terminating condition judgment operation, and outputting a final optimal individual as a cascade reservoir group flood control scheduling optimization result. According to the method, the novel flood optimization algorithm is introduced to solve the reservoir group optimization scheduling problem for the first time, the improved flood optimization algorithm based on the elite reverse strategy and the longitudinal crossing strategy is provided, and the method has the advantages of being high in convergence efficiency and high in calculation precision and has good popularization value and application prospects in solving the reservoir group optimization scheduling problem.

Description

Technical Field

[0001] The invention relates to the technical field of reservoir group optimization dispatching, and in particular to a flood control dispatching optimization method for a cascade reservoir group based on an improved flood optimization algorithm. Background Art

[0002] The optimal dispatching of flood control for cascade reservoirs is a typical high-dimensional, nonlinear, multi-stage, and strongly constrained complex system optimization and control problem. With the continuous development of cascade reservoirs in the basin, the scale of the main and tributary reservoirs is becoming increasingly large, and the contradiction between the complexity of the problem and the limitations of the solution method is more prominent, making the optimization solution very difficult. In order to solve this optimization problem, scholars at home and abroad have successively proposed different calculation methods. Although the traditional dynamic programming method can solve the optimal dispatching of a single reservoir, it faces the problem of dimensionality disaster when the research object is a reservoir group. Improved methods have also been proposed, such as discrete differential dynamic programming method and stepwise optimization method, which have alleviated the dimensionality disaster to a certain extent. In recent years, a series of intelligent evolutionary algorithms have also been proposed, adding new ideas to the optimal dispatching of reservoir groups, including particle swarm optimization, genetic algorithm, differential algorithm, quantum particle swarm optimization, etc., but different algorithms have their own advantages and limitations. There are problems such as difficulty in handling multiple constraints, long calculation time, and easy to fall into local optimality, which limits the practical application and promotion.

[0003] The Flood Algorithm (FLA) is a novel swarm intelligence optimization algorithm proposed in June 2024. The design of the algorithm is inspired by the study of the complex movement and flow patterns of floods. Its optimization idea is to simulate the evolution and changes of floods on the basis of population initialization, and to achieve iterative updates of the population to ultimately obtain the optimal result. The Flood Algorithm has the advantages of clear principles, simple structure, and fast optimization rate, but it is currently rarely used in the engineering field.

[0004] Therefore, the applicant considered proposing a new optimal scheduling method for cascade reservoirs based on a flood optimization algorithm. Summary of the invention

[0005] In order to overcome the shortcomings of the above-mentioned technology, the purpose of the present invention is to provide a flood control scheduling optimization method for cascade reservoir groups based on an improved flood optimization algorithm. On the basis of introducing the standard flood optimization algorithm into the flood control optimization scheduling of cascade reservoir groups, the standard flood optimization algorithm is improved based on elite reverse learning and vertical cross-learning strategies. It has the advantages of strong search capability and high calculation accuracy. It can quickly and accurately obtain the optimal scheduling plan of the cascade reservoir group and improve the optimization efficiency.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0007] A flood control dispatch optimization method for cascade reservoirs based on an improved flood optimization algorithm comprises the following steps:

[0008] 1) Obtain basic data on flood control and dispatching of cascade reservoirs;

[0009] 2) Determine the initial calculation conditions and establish the objective function of the optimal dispatch of cascade reservoirs for flood control as the fitness function;

[0010] 3) Population initialization operation, obtaining a population containing several individuals and obtaining the optimal individual in the population;

[0011] 4) Population optimization operation, which updates the population and the optimal individuals in the population through flood evolution simulation operation, flood change simulation operation, elite reverse learning operation, and vertical cross-learning operation;

[0012] 5) The termination condition judgment operation is performed and the final optimal individual is output as the flood control dispatch optimization plan for the cascade reservoir group.

[0013] Preferably, in step 2), the initial calculation conditions include taking the water level of each reservoir in each time period during the scheduling period as the decision variable, minimizing the sum of squares of the flow process of each flood control object as the optimization goal, establishing the objective function of the flood control optimization scheduling of the cascade reservoir group as the fitness function, and setting constraints.

[0014] Preferably, the constraints include water balance constraints of each reservoir, water level constraints of each reservoir, outflow constraints of each reservoir, and discharge capacity constraints of each reservoir.

[0015] Preferably, in step 3), the population initialization operation includes randomly initializing individuals according to the basic information and constraints of the cascade reservoirs to obtain a population containing a number of individuals and obtain the optimal individual of the population; including the following steps:

[0016] (1) Convert the water level of each reservoir in all periods to the individual variables of the population: Set the population size to NP and the maximum number of iterations to G, then the population consists of NP D-dimensional individuals, each of which is Where i = 1, 2, ..., NP, g = 0, 1, 2, ..., G, D = N × T;

[0017] Let the number of iterations g = 0, the initialized individual as follows:

[0018]

[0019] in, is the water level of the first reservoir at each time period, is the water level of the second reservoir at each time period, is the water level of the Nth reservoir at each time period; is the j-th dimension variable of individual i in the g-th generation population, i=1,2,…,NP; j=1,2,......,D, is the water level of the sth reservoir in the tth period, s=1,2,...,N; t=1,2,...,T;

[0020] (2) Obtain the optimal individual in the population: Select the individual with the smallest fitness value as the optimal individual in the population, denoted by x best =(x best,1 ,x best,2 ,...,x best,D ); the fitness value is obtained by calculating the fitness function f.

[0021] Preferably, in step 4), the flood evolution simulation operation, i.e., simulating conventional flood movement, comprises the following steps:

[0022] (1) Perform iterative update operations on all individuals to obtain mutant individuals The calculation formula is as follows:

[0023]

[0024] Among them, randn is a normally distributed random number; r1, r2, r3 and r4 are random numbers between 0 and 1; p j and q j are the upper and lower limits of the j-th dimension variable of the individual; is the j-th dimension variable of individual k randomly selected from the g-th generation population, k∈[1,NP] and k≠i; PK is the water consumption coefficient; PE i is the fitness coefficient value of individual i;

[0025] (2) Comparison of individuals and The fitness of Then Otherwise, no processing is done; where f is the fitness function;

[0026] (3) After all individuals have completed the flood evolution simulation operation, update the optimal individual x best :

[0027]

[0028] Preferably, the calculation formula of the water consumption coefficient PK is:

[0029]

[0030] Preferably, the fitness coefficient value PE of the individual i is i The calculation formula is:

[0031]

[0032] in,

[0033] Preferably, in step 4), the flood change simulation operation, i.e., simulating the evaporation attenuation of the flood, comprises the following steps:

[0034] (1) By sorting the updated population through the flood evolution simulation operation, NW poor individuals in the updated population are sorted. Perform iterative updates to obtain updated variant individuals The calculation formula is:

[0035]

[0036] Where r5 and r6 are random numbers between 0 and 1; x best,j The j-th dimension variable of the optimal individual after the flood evolution simulation operation is updated;

[0037] (2) Directly use the updated variant individuals Replace the worse individuals in the updated population That is

[0038] (3) Repeat steps (1) and (2) of the flood change simulation operation until all individuals have completed the flood change simulation operation and update the optimal individual x best .

[0039] Preferably, NW is 10% of the population number NP.

[0040] Preferably, in step 4), the elite reverse learning operation, i.e., introducing the elite reverse learning strategy on the basis of the standard flood optimization algorithm, comprises the following steps:

[0041] (1) Simulate the evolution of the flood and update the population with NE better elite individuals Perform iterative updates to obtain the updated reverse individual The calculation formula is:

[0042]

[0043] In the formula, r7 is a random number between 0 and 1;

[0044] (2) Comparison of individuals in the population after the flood evolution simulation operation update and the updated reverse individual The fitness of Then Otherwise no action will be taken.

[0045] Preferably, the value of NE is 10% of the population number NP.

[0046] Preferably, in step 4), the vertical cross-learning operation, i.e., introducing a vertical cross-learning strategy based on the standard flood optimization algorithm, comprises the following steps:

[0047] (1) Cross-learning is performed on the individual dimensional information of the group obtained through the flood evolution simulation operation, flood change simulation operation, and elite reverse learning operation. The j-dimensional and k-dimensional crossover individuals are updated The calculation formula is:

[0048]

[0049] Among them, r8 is a random number between 0 and 1; is the k-th dimension variable of the randomly selected individual i in the g-th generation population after being updated through the flood evolution simulation operation, flood change simulation operation, and elite reverse learning operation;

[0050] (2) Compare the individuals in the population obtained after updating the flood evolution simulation operation, flood change simulation operation, and elite reverse learning operation and the updated crossover individual The fitness of Then Otherwise, no processing will be done;

[0051] (3) Repeat steps (1) and (2) of the vertical cross-learning operation until all individuals have completed the vertical cross-learning operation and update the optimal individual x best .

[0052] Preferably, in step 5), the termination condition judgment operation is performed, that is, whether the current number of iterations reaches the maximum number of iterations is judged. If g≤G, set g=g+1 and repeat step 4); otherwise, stop the iteration and output the optimal individual as the final flood control scheduling optimization result of the cascade reservoir group.

[0053] The present invention also provides a cascade reservoir group flood control dispatch optimization system based on an improved flood optimization algorithm, which is used to implement the above-mentioned cascade reservoir group flood control dispatch optimization method based on the improved flood optimization algorithm, comprising:

[0054] The initial condition setting module is used to determine the objective function, constraints and decision variables for optimizing flood control in cascade reservoirs;

[0055] The population initialization operation module is used to randomly initialize individuals according to the constraints of the cascade reservoirs to obtain a population containing several individuals and compare and select the best individual;

[0056] Population optimization module, used to update the population and the optimal individuals in the population through flood evolution simulation operation, flood change simulation operation, elite reverse learning operation, and vertical cross-learning operation;

[0057] The output module is used to repeatedly execute the population optimization module until the iteration stop condition is met, obtain the optimal individual, and output it as the final optimization calculation result of the flood control dispatching water level of the cascade reservoir group.

[0058] Preferably, the population optimization module includes:

[0059] The flood evolution simulation operation submodule is used to simulate the regular flood movement and perform iterative updates on all individuals;

[0060] The flood change simulation operation submodule is used to simulate the evaporation attenuation of floods and iteratively update several poor individuals;

[0061] The elite reverse learning operation submodule is used to introduce the elite reverse learning strategy and iteratively update several elite individuals;

[0062] The vertical cross-learning operation submodule is used to introduce the vertical cross-learning strategy to perform cross-learning of all individuals' own dimensional information for iterative updating.

[0063] The present invention also provides a computer device, including a memory and a processor, wherein the memory is used to store at least one program, and the processor is used to load the at least one program to execute the above-mentioned flood control scheduling optimization method for cascade reservoirs based on the improved flood optimization algorithm.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] The flood optimization algorithm adopted in the present invention has the advantages of clear principle, simple structure and fast optimization rate. It mainly includes operations such as initialization, flood evolution simulation and flood change simulation, but it has not yet been applied to the field of reservoir group scheduling. This time, it is first introduced into the research problem of optimal scheduling of cascade reservoir groups.

[0066] Based on the standard flood optimization algorithm, the present invention incorporates elite reverse learning operations and vertical cross-learning operations, and proposes an improved flood optimization algorithm, which not only realizes the interactive learning of information between populations, and has the ability to develop and utilize information for global search and avoid local optimality, but also realizes the exchange and learning of information within individuals, and has the ability to deeply utilize information to jump out of local optimality and improve calculation accuracy.

[0067] The present invention discloses a flood control dispatch optimization method for cascade reservoirs based on an improved flood optimization algorithm, which has the advantages of simple principle, few control parameters, strong convergence ability, high calculation accuracy, and can be applied to solving complex optimization problems with multiple variables, multiple constraints, and high dimensions.

[0068] The present invention proposes a flood control scheduling optimization method for cascade reservoirs based on an improved flood optimization algorithm. The method has the advantages of strong search capability and effective avoidance of the "curse of dimensionality". It can quickly and effectively obtain an optimal scheduling plan for cascade reservoirs, and has good promotion value and application prospects in solving the optimal scheduling problem of reservoir groups. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a flow chart of a flood control dispatch optimization method of a cascade reservoir group based on an improved flood optimization algorithm of the present invention;

[0070] Figure 2 This is a diagram showing the calculation results of flood control and lake cruise dispatching of Reservoir A in the cascade reservoir group under the condition of 1% water inflow in 1998 in an embodiment of the present invention;

[0071] Figure 3 This is a diagram showing the calculation results of flood control optimization dispatching of Reservoir B in the cascade reservoir group under the 1% water inflow condition in 1998 in an embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to better explain the present invention, the main contents of the present invention are further explained below in conjunction with the drawings and specific embodiments, but the contents of the present invention are not limited to the following embodiments.

[0073] In order to reduce the flood control burden and flood losses downstream of the reservoir group and provide technical support for real-time flood control scheduling decisions, the maximum peak reduction criterion is currently widely used to construct the optimization scheduling objective function. To this end, based on the maximum peak reduction criterion, the cascade reservoir flood control optimization scheduling model is generally described as follows: given the initial water level, final water level, inflow flood process, and interval flood process of each reservoir during the scheduling period, while satisfying the complex constraints of the water level and flow of each reservoir and the upstream and downstream protection objects, the optimization goal is to minimize the sum of squares of the flow process of each flood control object. The specific mathematical expression of the flood control benefit target value is:

[0074]

[0075] Where: W is the square sum of the flow process of the flood control object; N is the number of reservoirs; i is the reservoir number, i = 1, 2, ..., N; T is the total number of time periods in the dispatch period; j is the time period number, j = 1, 2, ..., T; O i,j is the outflow of reservoir i in period j (m 3 / s); R i,j is the interval flow from reservoir i to the downstream reservoir or downstream flood control section in period j (m3 / s).

[0076] The constraints that need to be considered in the optimization calculation of flood control operation of cascade reservoirs mainly include:

[0077] (1) Water balance constraints of each reservoir:

[0078] (2) Hydraulic connection constraints of each reservoir:

[0079] (3) Water level constraints of each reservoir:

[0080] (4) Constraints on outflow from each reservoir:

[0081] (5) Constraints on discharge capacity of each reservoir:

[0082] (6) Initial and final water level constraints of each reservoir: Z i,0 =Z i,start ,Z i,T =Z i,end ;

[0083] (7) Non-negative constraint: All variables are non-negative values.

[0084] Where V i,j is the storage capacity of reservoir i at the end of period j (m 3 );I i,j is the inflow of reservoir i in period j (m 3 / s); i,j is the outflow of reservoir i in period j (m 3 / s); Δt is the time period length (h); and are the minimum outflow of reservoir i in period j (m 3 / s) and maximum outbound flow (m 3 / s); Z i,j is the water level in front of the dam of reservoir i in period j (m); and are the lowest and highest water levels in front of the dam of reservoir i in period j (m); O i,max (Z i,j ) is the water level Z corresponding to reservoir i in period j i,j Maximum discharge capacity (m 3 / s); Z i,start and Z i,end are the initial reservoir water level (m) and the controlled water level (m) at the end of the dispatching period of reservoir i, respectively.

[0085] By analyzing the above objective functions and constraints, it can be seen that the optimal scheduling of flood control for cascade reservoirs includes many complex constraints such as water level, storage capacity, flow rate, etc. It is a significant large-scale, high-dimensional, multi-stage, strongly constrained, nonlinear and other complex control problems.

[0086] In order to efficiently solve the flood control optimization scheduling problem of cascade reservoirs, the method of the present invention first analyzes the calculation principle of the flood optimization algorithm on the basis of introducing the standard flood optimization algorithm. On this basis, the elite reverse learning and vertical cross-learning strategies are further considered to make full use of the search information, and an improved flood optimization algorithm is proposed. The improved flood optimization algorithm has the advantages of strong search capability and effectively avoiding the "curse of dimensionality". It is a new method for efficiently solving the flood control optimization scheduling problem of cascade reservoirs.

[0087] like Figure 1 As shown, the present invention provides a flood control dispatch optimization method for cascade reservoirs based on an improved flood optimization algorithm, comprising the following steps:

[0088] 1) Obtain basic data on flood control and dispatching of cascade reservoirs;

[0089] Among them, the basic data include the number of reservoirs N, the number of scheduling periods T, the characteristic parameters of each reservoir, the water level and storage capacity curve of each reservoir, the discharge capacity curve of each reservoir, and the flood data of different times entering the reservoir.

[0090] 2) Determine the initial calculation conditions, including taking the water level of each reservoir in each period of the dispatch period as the decision variable, taking the minimum sum of squares of the flow process of each flood control object as the optimization goal, establishing the fitness function f, that is, the optimization dispatch objective function W mentioned above, and setting various constraints such as water level, flow, and reservoir capacity;

[0091] 3) Population initialization operation, including randomly initializing individuals according to the basic information and constraints of the cascade reservoirs to obtain a population containing several individuals. The population size is set to NP and the maximum number of iterations is G. Then the population consists of NP D-dimensional individuals, each of which is Where i = 1, 2, ..., NP, g = 0, 1, 2, ..., G, D = N × T; let the number of iterations g = 0, the individual initialization

[0092]

[0093] in, is the water level of the first reservoir at each time period, is the water level of the second reservoir at each time period, is the water level of the Nth reservoir at each time period; is the j-th dimension variable of individual i in the g-th generation population, i=1,2,…,NP; j=1,2,......,D, is the water level of the sth reservoir in the tth period, s = 1, 2, ..., N; t = 1, 2, ..., T. That is to say, one individual represents the water level of each reservoir in all periods, thus completing the coding conversion from water level to population individual variable.

[0094] At the same time, the individual with the smallest fitness value is selected as the optimal individual in the population, denoted by x best =(x best,1 ,x best,2 ,...,x best,D ). The fitness value is calculated by the fitness function f().

[0095] 4) Population optimization operation, which updates the population and the optimal individuals in the population through flood evolution simulation operation, flood change simulation operation, elite reverse learning operation, and vertical cross-learning operation.

[0096] 4.1) Flood evolution simulation operation, that is, simulating the movement of conventional floods

[0097] (1) Perform iterative update operations on all individuals to obtain mutant individuals The calculation formula is:

[0098]

[0099] Among them, randn is a normally distributed random number; r1, r2, r3 and r4 are random numbers between 0 and 1; p j and q j are the upper and lower limits of the reservoir water level of the individual j-th dimension variable; is the j-th dimension variable of the individual k randomly selected from the g-th generation population, with k∈[1,NP] and k≠i; PK is the water consumption coefficient, and the calculation formula is:

[0100]

[0101] PE i is the fitness coefficient value of individual i, and the calculation formula is:

[0102]

[0103] Among them, f() is the individual fitness calculation function;

[0104] (2) Comparison of individuals and The fitness of Then Otherwise no action will be taken.

[0105] (3) After all individuals have completed the flood evolution simulation operation, update the optimal individual xbest :

[0106]

[0107] 4.2) Flood change simulation operation, that is, simulating the evaporation attenuation of floods.

[0108] (1) By sorting the updated population through the flood evolution simulation operation, NW poor individuals in the updated population are sorted. Perform iterative updates, NW is generally taken as about 10% of NP, and the updated variant individuals are obtained The calculation formula is:

[0109]

[0110] (2) Directly use the updated variant individuals To replace the poor individuals in the updated population That is

[0111] (3) Repeat steps (1) and (2) of the flood change simulation operation until all individuals have completed the flood change simulation operation and update the optimal individual x best .

[0112] 4.3) Elite reverse learning operation, that is, introducing the elite reverse learning strategy based on the standard flood optimization algorithm.

[0113] (1) In addition to iteratively updating the poor individuals, we further consider the NE better elite individuals in the population after the flood evolution simulation operation is updated. Perform iterative updates, NE is generally taken as about 10% of NP, and the updated reverse individual The calculation formula is:

[0114]

[0115] Among them, r7 is a random number between 0 and 1.

[0116] (2) Comparison of individuals in the population after the flood evolution simulation operation update and the updated reverse individual The fitness of Then Otherwise no action will be taken.

[0117] 4.4) Vertical cross-learning operation, that is, on the basis of the standard flood optimization algorithm, in addition to exchanging information between individuals, a vertical cross-learning strategy is further introduced to consider cross-learning of information in various dimensions of the individual itself.

[0118] (1) The individual dimensional information of the individuals in the group obtained after updating through flood evolution simulation operation, flood change simulation operation and elite reverse learning operation is cross-learned based on The j- and k-dimensional crossover individuals The calculation formula is:

[0119]

[0120] Among them, r8 is a random number between 0 and 1; It is the k-th dimensional variable of the individual i randomly selected from the g-th generation population after being updated through the flood evolution simulation operation, the flood change simulation operation, and the elite reverse learning operation.

[0121] (2) Compare the individuals in the population obtained after updating the flood evolution simulation operation, flood change simulation operation, and elite reverse learning operation and the updated crossover individual The fitness of Then Otherwise no action will be taken.

[0122] (3) Repeat steps (1) and (2) of the vertical cross-learning operation until all individuals have completed the vertical cross-learning operation and update the optimal individual x best .

[0123] 5) Termination condition judgment operation, that is, judging whether the current number of iterations reaches the maximum number of iterations. If g≤G, let g=g+1 and repeat step 4); otherwise, stop the iteration and output the optimal individual as the final flood control dispatch optimization result of the cascade reservoir group.

[0124] The present invention also provides a cascade reservoir group flood control dispatch optimization system based on an improved flood optimization algorithm, which is used to implement the above-mentioned cascade reservoir group flood control dispatch optimization method based on the improved flood optimization algorithm, comprising:

[0125] The initial condition setting module is used to determine the objective function, constraints and decision variables for optimizing flood control in cascade reservoirs;

[0126] The population initialization operation module is used to randomly initialize individuals according to the constraints of the cascade reservoirs to obtain a population containing several individuals and compare and select the best individual;

[0127] The population optimization module is used to update the population and the optimal individuals in the population through flood evolution simulation operations, flood change simulation operations, elite reverse learning operations, and longitudinal cross-learning operations; the population optimization module includes: a flood evolution simulation operation submodule, which is used to simulate conventional flood movement and perform iterative updates on all individuals; a flood change simulation operation submodule, which is used to simulate the evaporation attenuation of floods and perform iterative updates on several poor individuals; an elite reverse learning operation submodule, which is used to introduce an elite reverse learning strategy and perform iterative updates on several elite individuals; and a longitudinal cross-learning operation submodule, which is used to introduce a longitudinal cross-learning strategy and perform cross-learning of all individuals' own dimensional information for iterative updates.

[0128] The output module is used to repeatedly execute the population optimization module until the iteration stop condition is met, obtain the optimal individual, and output it as the final optimization calculation result of the flood control dispatching water level of the cascade reservoir group.

[0129] The present invention also provides a computer device, including a memory and a processor, wherein the memory is used to store at least one program, and the processor is used to load the at least one program to execute the above-mentioned flood control scheduling optimization method for cascade reservoirs based on the improved flood optimization algorithm.

[0130] Next, taking three large cascade reservoirs on a tributary of a certain river in the upper reaches as an example, the technical solution of the present invention will be further explained.

[0131] Two flood control reservoirs, A and B, are built in the middle and lower reaches of the basin. Combined with the overall arrangement of the basin comprehensive planning and flood control planning, the cascade reservoirs form a whole to implement joint flood control scheduling, with a total flood control storage capacity of 2.5 billion m 3 , and plays a role in flood control in the basin. Taking the actual flood, 1% flood and 2% design flood in 1998 as an example, the flood control optimization dispatching of the cascade reservoir group was carried out, and the method of the present invention and POA were used to carry out the flood control optimization dispatching calculation of the cascade reservoir group. The specific calculation results are shown in Table 1.

[0132] Table 1: Comparison of calculation results of the method of the present invention and other optimization algorithms

[0133]

[0134]

[0135] From Table 1, we can see that for the water inflow process in three situations, the following findings are made:

[0136] (1) Compared with the calculation results of the standard flood optimization algorithm and the differential evolution algorithm, the results of the standard flood optimization algorithm are slightly worse. Although the standard flood optimization algorithm can be applied to the optimal scheduling of flood control of cascade reservoirs, its search ability needs to be improved and effective improvements are needed.

[0137] (2) Compared with the standard flood optimization algorithm and the differential evolution algorithm, the target value of the method of the present invention is smaller in the calculation results, indicating that the cascade reservoirs have played a good interception effect on the three floods, and also shows that the improvement strategy proposed this time is effective. The method of the present invention provides an efficient and practical way to deal with the optimization and dispatch of large-scale complex reservoir groups.

[0138] like Figures 2-3 The figure shows the calculation results of the flood control optimization dispatching of cascade reservoirs under the condition of 1% water inflow in 1998 by using the method of the present invention. The dispatching results all meet the constraints of water level, storage capacity, flow rate, etc. The outflow of the two reservoirs is relatively stable, indicating that the method of the present invention has excellent flood control, peak cutting and peak shifting effects. The method of the present invention can quickly obtain the optimized flood control dispatching plan of cascade reservoirs.

[0139] In summary, the method of the present invention is the first to introduce flood optimization methods into the optimization scheduling problem of cascade reservoirs. However, it is found that similar to the intelligent algorithm based on population iteration, its population diversity decreases in the later stage, and it is easy to fall into the local optimal defect, and the global search ability must be further improved. To this end, based on the analysis of the computational mechanism of the flood optimization algorithm, two improved operations are coupled: one is to propose elite reverse learning for the better individuals to increase the algorithm's ability to jump out of the local optimal, and the other is to propose learning the individual's own dimensional information and updating it by setting an effective information exchange mechanism to improve the algorithm's ability to avoid premature convergence.

[0140] The present invention proposes a flood control scheduling optimization method for cascade reservoirs based on an improved flood optimization algorithm. The method has the advantages of strong search capability and effective avoidance of the "curse of dimensionality". It can quickly and effectively obtain an optimal scheduling plan for cascade reservoirs, and has good promotion value and application prospects in solving the optimal scheduling problem of reservoir groups.

[0141] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A flood control dispatch optimization method for cascade reservoirs based on an improved flood optimization algorithm, characterized in that: The steps include: 1) Obtain basic data on flood control and dispatching of cascade reservoirs; 2) Determine the initial calculation conditions and establish the objective function of the optimal dispatch of cascade reservoirs for flood control as the fitness function; 3) Population initialization operation, obtaining a population containing several individuals and obtaining the optimal individual in the population; 4) Population optimization operation, which updates the population and the optimal individuals in the population through flood evolution simulation operation, flood change simulation operation, elite reverse learning operation, and vertical cross-learning operation; 5) The termination condition judgment operation is performed and the final optimal individual is output as the flood control dispatch optimization plan for the cascade reservoir group.

2. The method for optimizing flood control dispatching of cascade reservoirs according to claim 1 is characterized in that: In the step 2), the initial calculation conditions include taking the water level of each reservoir in each time period during the scheduling period as the decision variable, minimizing the sum of squares of the flow process of each flood control object as the optimization goal, establishing the objective function of the flood control optimization scheduling of the cascade reservoir group as the fitness function, and setting constraints.

3. The method for optimizing flood control dispatching of cascade reservoirs according to claim 2 is characterized in that: The constraints include water balance constraints of each reservoir, water level constraints of each reservoir, outflow constraints of each reservoir, and discharge capacity constraints of each reservoir.

4. The method for optimizing flood control dispatching of cascade reservoirs according to claim 2 is characterized in that: In the step 3), the population initialization operation includes randomly initializing individuals according to the basic information and constraints of the cascade reservoirs to obtain a population containing a number of individuals and obtain the optimal individual of the population; including the following steps: (1) Convert the water level of each reservoir in all periods to the individual variables of the population: Set the population size to NP and the maximum number of iterations to G, then the population consists of NP D-dimensional individuals, each of which is Where i = 1, 2, ..., NP, g = 0, 1, 2, ..., G, D = N × T; Let the number of iterations g = 0, the initialized individual as follows: in ……、 is the water level of the first reservoir at each time period, ……、 is the water level of the second reservoir at each time period, ..., ……、 is the water level of the Nth reservoir at each time period; is the j-th dimension variable of individual i in the g-th generation population, i=1,2,…,NP; j=1,2,......,D, is the water level of the sth reservoir in the tth period, s=1,2,...,N; t=1,2,...,T; (2) Obtain the optimal individual in the population: Select the individual with the smallest fitness value as the optimal individual in the population, denoted by x best =(x best,1 ,x best,2 ,...,x best,D ); the fitness value is obtained by calculating the fitness function f.

5. The method for optimizing flood control dispatching of cascade reservoirs according to claim 4 is characterized in that: In step 4), the flood evolution simulation operation, i.e., simulating the movement of conventional floods, includes the following steps: (1) Perform iterative update operations on all individuals to obtain mutant individuals The calculation formula is as follows: 1≤i≤NP,1≤j≤D Among them, randn is a normally distributed random number; r1, r2, r3 and r4 are random numbers between 0 and 1; p j and q j are the upper and lower limits of the j-th dimension variable of the individual; is the j-th dimension variable of individual k randomly selected from the g-th generation population, k∈[1,NP] and k≠i; PK is the water consumption coefficient; PE i is the fitness coefficient value of individual i; (2) Comparison of individuals and The fitness of Then Otherwise, no processing is done; where f is the fitness function; (3) After all individuals have completed the flood evolution simulation operation, update the optimal individual x best :

6. The method for optimizing flood control dispatching of cascade reservoirs according to claim 5 is characterized in that: The calculation formula of the water consumption coefficient PK is:

7. The method for optimizing flood control dispatching of cascade reservoirs according to claim 5 is characterized in that: The fitness coefficient value PE of the individual i i The calculation formula is: in, 8. The method for optimizing flood control dispatching of cascade reservoirs according to claim 5 is characterized by: In step 4), the flood change simulation operation, i.e., the simulation of the evaporation attenuation of the flood, comprises the following steps: (1) By sorting the updated population through the flood evolution simulation operation, NW poor individuals in the updated population are sorted. Perform iterative updates to obtain updated variant individuals The calculation formula is: Where r5 and r6 are random numbers between 0 and 1; x best,j The j-th dimension variable of the optimal individual after the flood evolution simulation operation is updated; (2) Directly use the updated variant individuals Replace the worse individuals in the updated population That is (3) Repeat steps (1) and (2) of the flood change simulation operation until all individuals have completed the flood change simulation operation and update the optimal individual x best .

9. The method for optimizing flood control dispatching of cascade reservoirs according to claim 8, characterized in that: The value of NW is 10% of the population size NP.

10. The method for optimizing flood control dispatching of cascade reservoirs according to claim 4, characterized in that: In step 4), the elite reverse learning operation, i.e., introducing the elite reverse learning strategy on the basis of the standard flood optimization algorithm, comprises the following steps: (1) Simulate the evolution of the flood and update the population with NE better elite individuals Perform iterative updates to obtain the updated reverse individual The calculation formula is: In the formula, r7 is a random number between 0 and 1; (2) Comparison of individuals in the population after the flood evolution simulation operation update and the updated reverse individual The fitness of Then Otherwise no action will be taken.

11. The method for optimizing flood control dispatching of cascade reservoirs according to claim 10, characterized in that: The value of NE is 10% of the population size NP.

12. The method for optimizing flood control dispatching of cascade reservoirs according to claim 4, characterized in that: In step 4), the vertical cross-learning operation, that is, introducing a vertical cross-learning strategy based on the standard flood optimization algorithm, comprises the following steps: (1) Cross-learning is performed on the individual dimensional information of the group obtained through the flood evolution simulation operation, flood change simulation operation, and elite reverse learning operation. The j- and k-dimensional crossover individuals are updated The calculation formula is: Among them, r8 is a random number between 0 and 1; is the k-th dimension variable of the randomly selected individual i in the g-th generation population after being updated through the flood evolution simulation operation, flood change simulation operation, and elite reverse learning operation; (2) Compare the individuals in the population obtained after updating the flood evolution simulation operation, flood change simulation operation, and elite reverse learning operation and the updated crossover individual The fitness of Then Otherwise, no processing will be done; (3) Repeat steps (1) and (2) of the vertical cross-learning operation until all individuals have completed the vertical cross-learning operation and update the optimal individual x best .

13. The method for optimizing flood control dispatching of cascade reservoirs according to claim 4, characterized in that: In the step 5), the termination condition judgment operation is performed, that is, whether the current number of iterations reaches the maximum number of iterations. If g≤G, set g=g+1 and repeat step 4); otherwise, stop the iteration and output the optimal individual as the final flood control scheduling optimization result of the cascade reservoir group.

14. A flood control dispatch optimization system for cascade reservoirs based on an improved flood optimization algorithm, characterized in that: The method for optimizing flood control dispatching of a cascade reservoir group based on an improved flood optimization algorithm for implementing any one of claims 1 to 13 comprises: The initial condition setting module is used to determine the objective function, constraints and decision variables for optimizing flood control in cascade reservoirs; The population initialization operation module is used to randomly initialize individuals according to the constraints of the cascade reservoirs to obtain a population containing several individuals and compare and select the best individual; Population optimization module, used to update the population and the optimal individuals in the population through flood evolution simulation operation, flood change simulation operation, elite reverse learning operation, and vertical cross-learning operation; The output module is used to repeatedly execute the population optimization module until the iteration stop condition is met, obtain the optimal individual, and output it as the final optimization calculation result of the flood control dispatching water level of the cascade reservoir group.

15. The flood control dispatch optimization system for cascade reservoirs according to claim 14 is characterized in that: The population optimization module includes: The flood evolution simulation operation submodule is used to simulate the regular flood movement and perform iterative updates on all individuals; The flood change simulation operation submodule is used to simulate the evaporation attenuation of floods and iteratively update several poor individuals; The elite reverse learning operation submodule is used to introduce the elite reverse learning strategy and iteratively update several elite individuals; The vertical cross-learning operation submodule is used to introduce the vertical cross-learning strategy to perform cross-learning of all individuals' own dimensional information for iterative updating.

16. A computer device, characterized in that: It comprises a memory and a processor, wherein the memory is used to store at least one program, and the processor is used to load the at least one program to execute the flood control scheduling optimization method for cascade reservoirs based on the improved flood optimization algorithm as described in any one of claims 1 to 13.

Images