Cascade reservoir dam front water level optimization scheduling method and program product

Abstract

The invention belongs to the technical field of reservoir optimization scheduling, and discloses a cascade reservoir pre-dam water level optimization scheduling method and a program product, and the method comprises the steps: S1, randomly generating a plurality of pre-dam water levels of a cascade reservoir as initial solutions; s2, calculating the generating capacity of each solution, and selecting the solutions corresponding to the first four maximum generating capacities; s3, selecting a reverse learning strategy or a variation strategy to execute disturbance through a preset switching probability P, generating a new solution, and updating according to a greedy mechanism; s4, randomly adopting a differential variation strategy or updating each solution according to a chimpanzee position updating formula fusing a water wave dynamic adaptive factor and a sine and cosine operator; s5, comparing the solutions before and after updating in the step S4; s6, judging whether an iteration termination condition is met or not, and if yes, outputting an optimal solution to obtain an optimal water level in front of the dam; otherwise, skipping to the step S2. According to the method, the convergence speed can be increased, premature convergence is effectively avoided, and the global search and local development capabilities of the original chimpanzee optimization algorithm are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of reservoir optimization and dispatching, and relates to an improved chimpanzee optimization algorithm integrating multiple strategies, and more specifically, to a method and program product for optimizing and dispatching water levels in front of dams of cascade reservoirs. Background Art

[0002] Reservoir optimization dispatching uses optimization technology to seek the optimal operation strategy that meets specific requirements such as power generation, flood control, and ecological flow. It is one of the main measures to reduce flood and drought disasters and achieve optimal allocation of water resources. With the completion and operation of large-scale water conservancy hubs, the intelligent management of reservoir group optimization dispatching has received more and more attention. However, due to the inherent multi-stage, multi-temporal and multi-constrained, and nonlinear characteristics, it is not an easy task to model and solve the optimal dispatching of cascade hydropower systems.

[0003] Although traditional optimization techniques such as linear programming, nonlinear programming, dynamic programming and its improved versions have been well applied in solving reservoir optimization scheduling problems, the efficiency of these methods is low, especially the dynamic programming algorithm, whose computational and storage capacity increases exponentially with the increase of the scale of cascade reservoirs, resulting in serious "curse of dimensionality" problems. In recent years, meta-heuristic optimization algorithms based on biological behavior or physical phenomena have gradually been favored by dispatchers, such as genetic algorithms (GA), differential evolution algorithms (DE), particle swarm optimization algorithms (PSO), and gray wolf optimization algorithms (GWO). Such algorithms can effectively overcome the shortcomings of traditional mathematical programming methods, and do not require the objective function and constraints to meet specific conditions. However, they have shortcomings such as premature convergence and poor robustness, which makes the optimal solution obtained when solving the optimization scheduling problem of cascade reservoirs often not ideal and difficult to guide the actual operation of the project. Therefore, the development of efficient optimization algorithms suitable for the optimization scheduling problem of cascade reservoirs is the key and difficult point of current research. Summary of the invention

[0004] In view of the above defects or improvement needs of the prior art, the present invention provides a method and program product for optimizing the water level in front of the dam of a cascade reservoir, which aims to solve the technical problem that the optimal solution is not ideal and it is difficult to guide the actual operation of the project.

[0005] To achieve the above object, according to one aspect of the present invention, a method for optimizing the water level in front of a cascade reservoir dam is provided, comprising the following steps:

[0006] S1, randomly generates multiple dam-front water levels of cascade reservoirs as initial solutions;

[0007] S2, compare the power generation of each solution, select the solutions corresponding to the top 4 largest power generation, and record them as the attacker chimpanzee X Attacker , Chimpanzee X Barrier , Chaser Chimpanzee XChaser and the chaser chimpanzee X Driver ;

[0008] S3, select the reverse learning strategy or the mutation strategy to perform perturbation on X with a preset switching probability P Attacker to generate a new solution X new,Attacker , and update X according to the greedy mechanism Attacker ;

[0009] S4, generate a random number u within a preset interval. If u≥δ, update each solution using the differential mutation strategy; otherwise, update each solution according to the chimpanzee position update formula that combines the water wave dynamic adaptive factor and the sine-cosine operator; δ is a preset value;

[0010] S5, compare the solutions before and after the update in step S4. If the generated power corresponding to the updated solution is larger, retain the updated solution; otherwise, randomly retain one of the solutions;

[0011] S6, determine whether the iteration termination condition is reached. If so, output the optimal solution X Attacker , and obtain the optimal water level in front of the dam; otherwise, jump to step S2.

[0012] Furthermore, the specific steps of S1 include:

[0013] S11, set the number of solutions NP, the maximum number of iterations T max , the dimension of the solution Dim, and the upper and lower boundaries Z up 、Z low ;

[0014] S12, randomly generate an initial solution according to the upper and lower boundaries Z up 、Z low of the water level in front of the dam.

[0015] Furthermore, the specific steps of S3 include:

[0016] S31, generate a switching probability P within the range of [0.6, 1.0], and generate a uniform random number r within the range of [0, 1];

[0017] S32, determine whether r<P is satisfied. If so, perform perturbation on X using the reverse learning strategy based on convex lens imaging Attacker to generate a new solution X new,Attacker , if not, generate X using the Cauchy mutation strategy new,Attacker ;

[0018] S33, update X according to the greedy mechanism Attacker .

[0019] Furthermore, the expression of the i-th solution vector in the t-th iteration in step S1 is as follows:

[0020]

[0021] in, represents the water level in front of the dam of the nth reservoir in the gth time period; n=1,2,…,N, g=1,2,…,G, N is the number of reservoirs, G is the number of time periods;

[0022] The calculation formula for randomly generating the initial solution in step S1 is as follows:

[0023]

[0024] Where r represents a uniform random number in the range [0,1]. Represents the upper and lower boundaries of the water level in front of the dam of the nth reservoir in the gth period.

[0025] Furthermore, step S3 adopts a reverse learning strategy based on convex lens imaging, and the calculation formula is:

[0026]

[0027] in, and They represent the attacker chimpanzee and its corresponding reverse solution in the t-th iteration respectively; η represents the scaling factor, which represents the proportional relationship between the object and the image, and its calculation formula is:

[0028]

[0029] Among them, η max and η min Represent the maximum and minimum scaling factors respectively; T max Indicates the maximum number of iterations.

[0030] Furthermore, the mutation strategy in step S3 is a Cauchy mutation strategy, and the calculation formula of the Cauchy mutation strategy is:

[0031]

[0032] in, represents the attacker chimpanzee in the tth iteration The solution after Cauchy mutation; cuachy(0,1) represents the Cauchy operator, that is, the standard Cauchy distribution; It means multiplication.

[0033] Furthermore, the greedy mechanism in step S3 is specifically as follows:

[0034]

[0035] in, Indicates that in the tth iteration, the attacker chimpanzee The new solutions generated; represents the attacker chimpanzee finally retained in the tth iteration; F(·) represents the power generation corresponding to the solution in the brackets.

[0036] Furthermore, the method for updating each solution according to the chimpanzee position update formula integrating the water wave dynamic adaptive factor and the sine-cosine operator described in step S4 is as follows:

[0037] Update the random sine convergence factor f according to the following formula:

[0038]

[0039] Among them, f max 、f min They represent the initial and final values ​​of the convergence factor respectively; rand represents a random number in the range of [0,1]; k represents the control factor, the larger its value, the slower the decrease speed of f;

[0040] Update coefficient vectors a, c, m:

[0041] a=2·f·r1-f

[0042] c=2·r2

[0043] Among them, r1 and r2 represent random vectors in the interval [0,1]; a represents a random vector that determines the distance between the solution and the optimal solution, and takes values ​​in the interval [-f,f]; c represents the control vector of the solution away from and close to the optimal solution, and takes values ​​in the interval [0,2]; m represents a chaotic mapping vector, which represents the influence of chimpanzee sexual motivation in the optimization process, and is calculated using Gauss / mouse map. The formula is as follows:

[0044]

[0045] Where t represents the number of iterations; the initial value of m is 0.7;

[0046] Chimpanzee Location Updates:

[0047]

[0048] Among them, X represents the solution in the current iteration; X(t+1) represents the updated solution in the current iteration; λ represents the water wave dynamic adaptive factor; s1·sin(s2) represents the sine operator; s1·cos(s2) represents the cosine operator; s1 represents the key parameter of the global exploration and local development capabilities of the balance algorithm; s2 represents a random number in the interval [0,2π]; s3 represents a random number in the interval [0,1]; α represents a constant.

[0049] Furthermore, in step S5, the Metropolis criterion is used to compare the solutions before and after the update, as follows:

[0050] Assume that at temperature T, the new solution generated by the current solution i is j, and the power generation corresponding to the two solutions is E i and E j , if E i <E j , then accept the new solution j, otherwise determine whether to keep the new solution j according to the probability w that the system is in the new solution j. The expression of probability w is:

[0051]

[0052] Among them, K B represents the Boltzmann constant; T(t) represents the cooling function, whose value decreases with the increase of the iteration number t;

[0053] Generate a random number ξ in the interval [0,1]. If w>ξ, accept the new solution j, otherwise keep the solution i.

[0054] According to another aspect of the present invention, a program product for optimizing the scheduling of water levels in front of dams of cascade reservoirs is provided. When the program product is called and executed, the method for optimizing the scheduling of water levels in front of dams of cascade reservoirs as described in any of the preceding items is implemented.

[0055] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0056] The proposed method for optimizing the water level in front of the dam of cascade reservoirs is improved by integrating multiple strategies. Among them, the global exploration and local development capabilities can be balanced by comparing and updating, and the sine and cosine operators. The reverse learning strategy, mutation strategy and water wave dynamic adaptive factor can enhance the diversity of the population and also help the algorithm to jump out of the local optimum. Therefore, the present invention improves the convergence and robustness of the optimization process, making the optimization and scheduling results of the water level in front of the dam of cascade reservoirs more practical in engineering. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 The present invention is a flow chart of a method for optimizing the water level in front of a cascade reservoir dam according to a preferred embodiment of the present invention.

[0058] Figure 2 It is a schematic diagram of the location of the Three Gorges-Gezhouba cascade system in the middle and lower reaches of the Yangtze River in a preferred embodiment of the present invention.

[0059] Figure 3 This is a convergence trajectory diagram obtained by each method under different typical water inflow conditions in the preferred embodiment of the present invention.

[0060] Figure 4Box plots obtained by various methods under different typical water inflow conditions in the preferred embodiment of the present invention.

[0061] FIG. 5( a ) and FIG. 5( b ) are diagrams showing the water level processes of cascade reservoirs obtained by various methods under different typical water inflow conditions in a preferred embodiment of the present invention.

[0062] Figure 6(a) to Figure 6(c) This is a process diagram of power generation of cascade reservoirs obtained by several methods under different typical water inflow conditions in the preferred embodiment of the present invention. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0064] like Figure 1 The flowchart of a preferred method for optimizing the water level in front of a cascade reservoir dam of the present invention is shown, which includes the following steps:

[0065] S1, set relevant parameters and randomly generate the initial solution, i.e., the water level in front of the dam of the cascade reservoir;

[0066] S2, calculate the power generation of each solution by combining the constraint processing method, select the solutions corresponding to the top 4 largest power generation, and record them as the attacker chimpanzee X Attacker , Chimpanzee X Barrier , Chaser Chimpanzee X Chaser and the drive-away chimpanzee X Driver ;

[0067] S3, through the switching probability P between [0.6, 1.0], select the reverse learning strategy based on convex lens imaging or the Cauchy mutation strategy for X Attacker Execute the perturbation to generate a new solution X new,Attacker , and update X according to the greedy mechanism Attacker ;

[0068] S4, generate a random number u in the interval [0,1]. If u ≥ 0.5, use the differential mutation strategy "rand / 1" to update each solution. Otherwise, update each solution according to the chimpanzee position update formula that integrates the water wave dynamic adaptive factor and the sine-cosine operator.

[0069] S5, comparing the solutions before and after the update in step S5, if the updated solution corresponds to a larger power generation, retaining the updated solution; otherwise, randomly retaining one of the solutions;

[0070] S6, Determine whether the iteration termination condition is reached. If so, end the program and output the optimal solution X Attacker , otherwise, jump to step S2.

[0071] In this embodiment, step S1 specifically includes:

[0072] S11, Set the number of solutions NP, the maximum number of iterations T max , the dimension of the solution Dim, the upper and lower boundaries of the water level in front of the dam Z up 、Z low , the initial temperature T0, the maximum and minimum scaling factors η max 、η min , the initial and final values f of the random sine convergence factor max 、f min , the control factor k, the differential ratio factor F, the constant α, the cooling function T(t) and the chaotic mapping vector m;

[0073] S12, According to the upper and lower boundaries of the water level in front of the dam Z up 、Z low , randomly generate the initial solution.

[0074] In this embodiment, step S3 specifically includes,

[0075] S31, Generate the switching probability P within the range of [0.6, 1.0], and generate a uniform random number r within the range of [0, 1];

[0076] S32, Determine whether r < P is satisfied. If so, perform perturbation on X using the reverse learning strategy based on convex lens imaging Attacker to generate a new solution X new,Attacker， If not, generate X using the Cauchy mutation strategy new,Attacker ;

[0077] S33, Update X according to the greedy mechanism Attacker .

[0078] In this embodiment, the expression of the i-th solution vector in the t-th iteration in step S1 is as follows:

[0079]

[0080] Where, represents the water level in front of the dam of the n-th reservoir at the g-th time period, n = 1, 2, …, N, g = 1, 2, …, G, N is the number of reservoirs, and G is the number of time periods.

[0081] In this embodiment, the calculation formula of the random initial solution in step S1 is as follows:

[0082]

[0083] Where r represents a uniform random number in the range [0,1]. Represents the upper and lower boundaries of the water level in front of the dam of the nth reservoir in the gth period.

[0084] In this embodiment, the constraint processing method in step S2 is as follows:

[0085] First, the infeasible solution generated in each iteration is dynamically adjusted to the upper and lower boundaries of the water level in front of the dam through boundary constraints. On this basis, the penalty function method is used to further deal with the violation of constraints. The calculation formula is as follows:

[0086]

[0087] in, represents the i-th solution in the t-th iteration Power generation; n,t represents the constraint violation value of the nth reservoir in the gth period; and c a represents the ath inequality constraint and its penalty coefficient; and c b represents the bth equality constraint and its penalty coefficient; A and B represent the number of inequality constraints and equality constraints, respectively.

[0088] In this embodiment, the calculation formula of the reverse learning strategy based on convex lens imaging in step S3 is as follows:

[0089]

[0090] in, and They represent the attacker chimpanzee and its corresponding reverse solution in the tth iteration respectively; η represents the scaling factor, representing the proportional relationship between the object and the image. Generally, a fixed value is used. In order to balance the global search and local development capabilities of the algorithm, a nonlinear dynamic scaling factor is used here, and its calculation formula is:

[0091]

[0092] Among them, η max and η min Represent the maximum and minimum scaling factors respectively; T max Indicates the maximum number of iterations.

[0093] In this embodiment, the calculation formula of the Cauchy mutation strategy in step S3 is as follows:

[0094]

[0095] in, represents the attacker chimpanzee in the tth iteration The solution after Cauchy mutation; cuachy(0,1) represents the Cauchy operator, that is, the standard Cauchy distribution; It means multiplication.

[0096] In this embodiment, the greedy mechanism in step S3 is as follows:

[0097]

[0098] in, Indicates that in the tth iteration, the attacker chimpanzee The new solutions generated; represents the attacker chimpanzee finally retained in the tth iteration; F(·) represents the power generation corresponding to the solution in the brackets.

[0099] In this embodiment, the calculation formula of the random sine convergence factor f in step S4 is as follows:

[0100]

[0101] Among them, f max 、f min They represent the initial and final values ​​of the convergence factor respectively; rand represents a random number in the range of [0,1]; k represents the control factor. The larger its value is, the slower the rate of decrease of f is. After many experiments, it is found that the best effect is achieved when k is 2.5.

[0102] In this embodiment, the calculation formulas of the coefficient vectors a, c, and m in step S4 are as follows:

[0103] a=2·f·r1-f

[0104] c=2·r2

[0105] m=Chaotic_value

[0106] Among them, r1 and r2 represent random vectors in the interval [0,1]; a represents a random vector that determines the distance between the solution and the optimal solution, and takes values ​​in the interval [-f,f]; c represents the control vector of the solution away from and close to the optimal solution, and takes values ​​in the interval [0,2]; m represents a chaotic mapping vector, which represents the influence of chimpanzee sexual motivation in the optimization process, and is calculated using Gauss / mouse map. The formula is as follows:

[0107]

[0108] Where t represents the number of iterations; the initial value of m is 0.7.

[0109] In this embodiment, the differential mutation strategy "rand / 1" is adopted in step S4, and the calculation formula is as follows:

[0110]

[0111] Where F represents the differential scale factor; represents the i-th solution in the t-th iteration; p1, p2, p3 represent randomly generated different integers in the range [1, NP], where NP is the number of solutions; V i t represents the i-th variation solution in the t-th iteration.

[0112] In this embodiment, the chimpanzee position update formula integrating the water wave dynamic adaptive factor and the sine-cosine operator in step S5 is as follows:

[0113]

[0114] Among them, X represents the solution in the current iteration; X(t+1) represents the updated solution in the current iteration; λ represents the water wave dynamic adaptive factor; s1·sin(s2) represents the sine operator; s1·cos(s2) represents the cosine operator; s1 represents the key parameter of the global exploration and local development capabilities of the balance algorithm; s2 represents a random number in the interval [0,2π]; s3 represents a random number in the interval [0,1]; α represents a constant.

[0115] In this embodiment, the Metropolis criterion in step S5 is as follows:

[0116] For the optimal operation of cascade reservoirs, the Metropolis criterion assumes that at temperature T, the new solution generated by the current solution i is j, and the power generation corresponding to the two solutions is E i and E j , if E i <E j , then accept the new solution j, otherwise judge whether the solution is an "important" solution based on the probability that the system is in the new solution j. The expression of this probability is:

[0117]

[0118] Among them, K B represents the Boltzmann constant, which takes a value of 1; T(t) represents the cooling function, which decreases with the increase of the number of iterations t, and controls the global search and local development capabilities of the algorithm. Generate a random number ξ in the interval [0,1]. If w>ξ, then the new solution j is an "important" solution and is accepted. Otherwise, solution i is still retained.

[0119] In this embodiment, the updating formula of the temperature T in step S5 is as follows:

[0120] T(t)=0.9×T(t-1)

[0121] Where t is the current iteration number.

[0122] In this embodiment, the iteration termination condition in step S6 is reaching the maximum number of iterations.

[0123] Example: Taking the Three Gorges-Gezhouba cascade system in the middle and lower reaches of the Yangtze River as an example, its geographical location is as follows Figure 2 As shown in the figure, based on the hydrological frequency analysis of the water inflow sequence data at the ten-day scale from 2004 to 2021, three typical years of relatively abundant (P = 25%), average (P = 50%), and relatively dry (P = 75%) are selected for power generation optimization scheduling calculation, and the scheduling results of the method of the present invention and five methods (DE, WOA, ALO, FDA, CHOA) are compared and analyzed to verify the superiority of the performance of the present invention in the optimal scheduling of cascade reservoirs. The objective function is to maximize the total power generation of the cascade reservoirs, the scheduling period is set to 1 year, the scheduling step is ten days, the number of solutions is 50, the maximum number of iterations is 500, and the values ​​of the other parameters are shown in Table 1. To avoid the adverse effects of random seeds, each algorithm is run independently 10 times.

[0124] Table 1 Algorithm parameter settings

[0125]

[0126] Table 2 gives the statistical results of the power generation optimization scheduling calculation of each method under different typical water inflow conditions. The optimal results in the table are expressed in bold. It can be seen that for each level year runoff, the present invention (MS-ICHOA) has obtained the best results in the four indicators of optimal value, average value, worst value and standard deviation. For example, compared with DE, WOA, ALO, FDA and CHOA, the optimal power generation of the present invention under the condition of abundant water inflow is increased by 0.41%, 0.23%, 0.11%, 0.33% and 0.87% respectively, the average power generation under the condition of normal water inflow is increased by 1.60%, 0.26%, 0.16%, 1.95% and 0.30% respectively, and the standard deviation under the condition of low water inflow is reduced by 98.42%, 98.46%, 31.17%, 99.35% and 97.85% respectively, indicating that the proposed method can effectively improve the total power generation of cascade reservoirs, increase power generation efficiency, and has strong stability.

[0127] Table 2 Statistical results of power generation optimization dispatch calculations by various methods under different typical water inflow conditions

[0128]

[0129] Figure 3Represents the comparison between the average convergence trajectories obtained by each method. For relatively abundant, average, and relatively low water conditions, the present invention rapidly improves the quality of the solution as the number of iterations increases in the early stages of iteration, and slowly approaches the optimal solution after the number of iterations reaches about 20. Although the convergence speeds of WOA and ALO in the early stages are comparable to those of the present invention, especially ALO found a better solution at the beginning of the iteration, the optimal power generation ultimately obtained is not as high as that of the present invention. In contrast, the number of iterations required by the other algorithms to find a feasible solution is relatively large. Therefore, the present invention not only has a fast convergence speed when solving the problem of optimal scheduling of cascade reservoirs, but also has strong global search capabilities and local development capabilities, and can effectively avoid falling into local optimality.

[0130] Figure 4 The box plots of the optimization results of the six methods under various typical water inflow conditions are depicted. It is found that the optimal solution obtained by the present invention is the highest and has the smallest distribution range among all the methods, which once again shows that the present invention has strong robustness and is a powerful tool for dealing with the problem of optimal scheduling of cascade reservoirs.

[0131] Figure 5(a) to Figure 5(b) The water level change process of the cascade reservoirs represents the optimal scheduling results obtained by several representative algorithms under various typical water inflow scenarios. It can be seen that for all methods, the water levels of the Three Gorges Reservoir and the Gezhouba Reservoir are within their water level constraints during the scheduling period. During the flood season, the Three Gorges Reservoir is maintained at a lower water level to leave enough storage capacity to prevent floods and ensure flood control safety in the downstream area; during the non-flood season, compared with DE, the water level process of the Three Gorges Reservoir obtained by the present invention enables the reservoir to always operate at a high water level, while increasing power generation, providing compensation benefits for other reservoirs in the basin, and ensuring water for life, production, ecology and agriculture. In addition, the water level operation process of the Gezhouba Reservoir obtained by the present invention can reduce the fluctuation of the water level in front of the dam and increase safety. It is worth noting that in all typical cases or some of them, the water level change trends of WOA, ALO, FDA and CHOA are similar to those of the present invention, but these methods are easily affected by random initial solutions, and the results of re-running may not be the case, so the overall effect is not as good as the present invention.

[0132] Figure 6(a) to Figure 6(c)The figure shows the change process of power generation of cascade reservoirs with the optimal scheduling results obtained by several representative algorithms under various typical water inflow scenarios. It is found that the power generation of the two reservoirs is relatively large during the flood season. The difference from the results of DE and CHOA is that, firstly, during the 5th to 12th decade, the power generation of the present invention is generally lower than that of DE, which corresponds to the higher water level obtained by the present invention during this period; secondly, before the flood season, the power generation of the present invention increases significantly and does not produce a sudden change like CHOA, which is due to the fact that the Three Gorges Reservoir increases the discharge flow to prepare for flood control during the flood season. Although DE and CHOA can eventually drop to the flood control limit water level before the flood season, it is obvious that the method of the present invention significantly improves the power generation of the entire scheduling period during this period, which can also be seen from Figure 6a )~Figure 6(c) shows the cumulative power generation change process of single reservoir and cascade reservoir. This case analysis fully proves the applicability and superiority of the present invention in the problem of reservoir optimization scheduling.

[0133] In summary, when solving complex engineering problems such as the optimal scheduling of cascade reservoirs, compared with several existing mature algorithms, the present invention can demonstrate superior performance such as fast convergence speed, good robustness, and effective avoidance of "premature convergence", and can significantly increase power generation while producing a more reasonable scheduling process, indicating that the improved chimpanzee optimization algorithm (MS-ICHOA) that integrates multiple strategies proposed in the present invention is an efficient method for solving the optimal scheduling problem of cascade reservoirs.

[0134] 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 method for optimizing the water level in front of a cascade reservoir dam, characterized in that: The steps include: S1, randomly generates multiple dam-front water levels of cascade reservoirs as initial solutions; S2, compare the power generation of each solution, select the solutions corresponding to the top 4 largest power generation, and record them as the attacker chimpanzee X Attacker , Chimpanzee X Barrier , Chaser Chimpanzee X Chaser and the evictor chimpanzee X Driver ; S3, select the reverse learning strategy or mutation strategy for X through the preset switching probability P Attacker Execute the perturbation to generate a new solution X new,Attacker , and update X according to the greedy mechanism Attacker ; S4, generating a random number u within a preset interval, if u ≥ δ, using the differential mutation strategy to update each solution, otherwise updating each solution according to the chimpanzee position update formula integrating the water wave dynamic adaptive factor and the sine-cosine operator; δ is a preset value; S5, comparing the solutions before and after the update in step S4, if the updated solution corresponds to a larger power generation, retain the updated solution; otherwise, randomly retain one of the solutions; S6, determine whether the iteration termination condition is reached, if so, output the optimal solution X Attacker , get the optimal water level in front of the dam; otherwise jump to step S2.

2. The method for optimizing the water level in front of a cascade reservoir dam according to claim 1, characterized in that: The S1 specifically includes: S11, set the number of solutions NP, the maximum number of iterations T max , the dimension of the solution Dim, the upper and lower boundaries of the water level in front of the dam Z up , Z low ; S12, according to the upper and lower boundaries of the water level in front of the dam Z up , Z low , randomly generate initial solutions.

3. The method for optimizing the water level in front of a cascade reservoir dam according to claim 2, characterized in that: The S3 specifically includes: S31, generate a switching probability P in the range of [0.6, 1.0] and a uniform random number r in the range of [0, 1]; S32, determine whether r < P is satisfied. If so, adopt the reverse learning strategy based on convex lens imaging to perform perturbation on X Attacker to generate a new solution X new,Attacker , if not, adopt the Cauchy mutation strategy to generate X new,Attacker ; S33, update X according to the greedy mechanism Attacker .

4. The method for optimizing the water level in front of a cascade reservoir dam according to claim 3, characterized in that: The i-th solution vector in the t-th iteration in step S1 The expression is as follows: in, represents the water level in front of the dam of the nth reservoir in the gth time period; n=1,2,…,N, g=1,2,…,G, N is the number of reservoirs, G is the number of time periods; The calculation formula for randomly generating the initial solution in step S1 is as follows: Where r represents a uniform random number in the range [0,1]. Represents the upper and lower boundaries of the water level in front of the dam of the nth reservoir in the gth period.

5. The method for optimizing the water level in front of a cascade reservoir dam according to claim 4, characterized in that: Step S3 adopts a reverse learning strategy based on convex lens imaging, and the calculation formula is: in, and They represent the attacker chimpanzee and its corresponding reverse solution in the t-th iteration respectively; η represents the scaling factor, which represents the proportional relationship between the object and the image, and its calculation formula is: Among them, η max and η min Respectively represent the maximum and minimum scaling factors; T max Indicates the maximum number of iterations.

6. A method for optimizing the water level in front of a cascade reservoir dam according to claim 5, characterized in that: The mutation strategy in step S3 is the Cauchy mutation strategy, and the calculation formula of the Cauchy mutation strategy is: in, represents the attacker chimpanzee in the tth iteration The solution after Cauchy mutation; cuachy(0,1) represents the Cauchy operator, that is, the standard Cauchy distribution; It means multiplication.

7. A method for optimizing the water level in front of a cascade reservoir dam according to claim 6, characterized in that: The greedy mechanism described in step S3 is as follows: in, Indicates that in the tth iteration, the attacker chimpanzee The new solutions generated; represents the attacker chimpanzee finally retained in the tth iteration; F(·) represents the power generation corresponding to the solution in the brackets.

8. A method for optimizing the water level in front of a cascade reservoir dam according to claim 7, characterized in that: The method for updating each solution according to the chimpanzee position update formula integrating the water wave dynamic adaptive factor and the sine-cosine operator described in step S4 is as follows: Update the random sine convergence factor f according to the following formula: Among them, f max 、f min They represent the initial and final values ​​of the convergence factor respectively; rand represents a random number in the range of [0,1]; k represents the control factor, the larger its value, the slower the decrease speed of f; Update coefficient vectors a, c, m: a=2·f·r1-f c＝2·r2 Among them, r1 and r2 represent random vectors in the interval [0,1]; a represents a random vector that determines the distance between the solution and the optimal solution, and takes values ​​in the interval [-f,f]; c represents the control vector of the solution away from and close to the optimal solution, and takes values ​​in the interval [0,2]; m represents a chaotic mapping vector, which represents the influence of chimpanzee sexual motivation in the optimization process, and is calculated using Gauss / mouse map. The formula is as follows: Where t represents the number of iterations; the initial value of m is 0.7; Chimpanzee Location Updates: Among them, X represents the solution in the current iteration; X(t+1) represents the updated solution in the current iteration; λ represents the water wave dynamic adaptive factor; s1·sin(s2) represents the sine operator; s1·cos(s2) represents the cosine operator; s1 represents the key parameter of the global exploration and local development capabilities of the balance algorithm; s2 represents a random number in the interval [0,2π]; s3 represents a random number in the interval [0,1]; α represents a constant.

9. A method for optimizing water level in front of dams of cascade reservoirs according to claim 8, characterized in that: In step S5, the Metropolis criterion is used to compare the solutions before and after the update, as follows: Assume that at temperature T, the new solution generated by the current solution i is j, and the power generation corresponding to the two solutions is E i and E j , if E i <E j , then accept the new solution j, otherwise determine whether to keep the new solution j according to the probability w that the system is in the new solution j. The expression of probability w is: Among them, K B represents the Boltzmann constant; T(t) represents the cooling function, whose value decreases with the increase of the iteration number t; Generate a random number ξ in the interval [0,1]. If w>ξ, accept the new solution j, otherwise keep the solution i.

10. A cascade reservoir dam front water level optimization scheduling program product, characterized in that: When the program product is called and executed, the method for optimizing the water level in front of the dam of the cascade reservoirs as described in any one of claims 1 to 9 is implemented.

Images