A general optimization method and system for water supply scheduling of parallel reservoir groups

Abstract

This invention provides a general optimization method and system for water supply scheduling of parallel reservoir groups. By acquiring system operation data of the parallel reservoir group and its water supply network, a multi-level, multi-attribute generalized directed graph model is constructed. Based on this model, the dynamic operating state of the water resource network is dynamically generalized, and multiple objective functions and constraints are dynamically configured. Simultaneously, scheduling decision variables are identified and the degree of constraint violation is quantified, forming a generalized optimization problem definition. Different optimization algorithms are supported, and the optimization problem is solved in multiple cycles through a scheduling decision evaluation module. The TOPSIS method is used to select representative solutions, obtaining the optimal release decision for each reservoir in each scheduling cycle. Finally, the performance of the scheduling scheme is evaluated based on simulation results, and a detailed operation report is generated. This invention can effectively improve the scientific and economic efficiency of water supply scheduling of parallel reservoir groups, providing important technical support for regional water resource optimization scheduling and refined management.

Description

Technical Field

[0001] This invention relates to the field of water resource scheduling technology, and in particular to a universal optimization method and system for water supply scheduling of parallel reservoir groups. Background Technology

[0002] The global water shortage and uneven distribution are becoming increasingly severe, making the optimal allocation and efficient utilization of water resources a key issue in water resource management. As a crucial component of regional water supply security, the coordinated operation of parallel reservoir groups directly impacts regional water security, economic development, and ecological balance. However, against the backdrop of frequent extreme hydrological events due to climate change and continuously increasing water demand driven by socio-economic development, the operation of parallel reservoir groups faces increasingly complex challenges.

[0003] Currently, various optimization algorithms, including mathematical programming and intelligent optimization algorithms, are widely used in reservoir water supply scheduling problems, combined with multi-objective optimization techniques to improve scheduling efficiency. However, existing technologies still have many shortcomings in handling the water supply scheduling problem of parallel reservoir groups. For example, most existing scheduling models cannot efficiently and accurately represent the complex topology of parallel reservoir groups, including multiple water sources, multiple water demand points, multiple water transmission links, and confluence / diversion points. Specifically, existing models often use single node types and fixed link attributes, making it difficult to generalize heterogeneous physical entities such as reservoirs, water users, and confluence / diversion points. Furthermore, they cannot uniformly integrate multi-dimensional, time-varying parameters such as monthly dynamically changing capacity or flow limits, interval inflow and loss, water loss rate, and water storage priority weights, severely limiting the model's ability to capture the dynamic behavior of real-world complex systems and its versatility. In addition, for different scheduling cases, existing models often tightly couple and hard-code the optimization objective function and constraints with specific network structures or scheduling logic, resulting in poor versatility, high development and maintenance costs, and difficulty in quickly adapting to diverse scheduling scenarios or system structure changes. Meanwhile, existing models are often tightly bound to specific optimization algorithms, making it difficult for users to flexibly switch between and compare the performance of different algorithms according to specific problem requirements. Furthermore, they lack general and advanced decision-making methods for selecting representative solutions in multi-objective optimization, which may affect the exploration of the optimality of scheduling schemes.

[0004] Overall, there are few records in domestic and foreign literature on generalized solution methods applicable to water supply scheduling of parallel reservoir groups, especially no reports on generalized optimization solution methods that integrate multi-level, multi-attribute generalized topology modeling, accurate generalization of dynamic operating states, and representative solution selection strategies. Summary of the Invention

[0005] This invention provides a general optimization method and system for water supply scheduling of parallel reservoir groups, which solves the shortcomings of existing technologies in dealing with water supply scheduling of complex parallel reservoir groups, such as insufficient model topology representation capability, poor versatility leading to high customized development costs, and inconvenience in switching optimization algorithms.

[0006] In a first aspect, the present invention provides a universal optimization method for water supply scheduling of a parallel reservoir group, comprising: Obtain system operation data of each reservoir in a parallel reservoir group, and construct a generalized directed graph model based on the system operation data; Based on the generalized generalized directed graph model, the dynamic operation state of the parallel reservoir group water resource network is generalized, the directed edge flow matrix and node storage matrix are determined, and the dynamic relationship between the matrix elements is deduced through topological sorting and general water balance equation. Based on the actual needs of water supply scheduling of parallel reservoir groups, the objective function and constraints are dynamically selected and parameterized for optimization, the scheduling decision variable matrix is ​​identified, and the degree of violation of each constraint is quantified to form a general optimization problem definition. A parallel reservoir group scheduling decision evaluation module is constructed, which receives the scheduling decision variable matrix, outputs the dynamic operation state set of the network through physical simulation calculation, and performs iterative solution based on the optimization algorithm module to obtain the optimal flow allocation scheme within the optimization period. By selecting a representative solution using TOPSIS, the system state driven by the optimal solution in the previous optimization cycle is used as the initial condition for the current optimization cycle. Multi-cycle rolling optimization is performed until the entire simulation period is completed, and the simulation results during the entire simulation period are recorded to generate a detailed operation report.

[0007] Secondly, the present invention also provides a universal optimization system for water supply scheduling of parallel reservoir groups, comprising: The data input module is used to acquire system operation data of each reservoir in the parallel reservoir group and to construct a generalized directed graph model based on the system operation data. The topology construction module is used to generalize the dynamic operation state of the parallel reservoir group water resources network based on the generalized generalized model of the generalized directed graph, determine the directed edge flow matrix and the node water storage matrix, and deduce the dynamic relationship between the matrix elements through topological sorting and generalized water balance equation. The problem assessment module is used to dynamically select and parameterize the objective function and constraints for optimization based on the actual needs of water supply scheduling of the parallel reservoir group, identify the scheduling decision variable matrix, quantify the degree of violation of each constraint, and form a general optimization problem definition. The optimization solution module is used to construct the parallel reservoir group scheduling decision evaluation module. It receives the scheduling decision variable matrix, outputs the network dynamic operation state set through physical simulation calculation, and performs iterative solution based on the optimization algorithm module to obtain the optimal flow allocation scheme within the optimization period. The results reporting module is used to select representative solutions through TOPSIS, take the system state driven by the optimal solution of the previous optimization cycle as the initial condition of the current optimization cycle, perform multi-cycle rolling optimization solutions until the entire simulation period is completed, record the simulation results during the entire simulation period, and generate a detailed operation report.

[0008] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the general optimization method for water supply scheduling of parallel reservoir groups as described above.

[0009] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the general optimization method for water supply scheduling of parallel reservoir groups as described above.

[0010] The universal optimization method and system for water supply scheduling of parallel reservoir groups provided by this invention achieves effective decoupling and unified integration of data input, topology construction, optimization problem definition, physical simulation, and optimization solution through modular design. By establishing a multi-level, multi-attribute generalized directed graph model, combined with dynamically configured optimization objectives and constraints, and flexibly integrated multiple optimization algorithms, the model can adapt to different scheduling scenarios and system structures, significantly reducing the cost of customized development and maintenance. It provides important and highly operable scientific decision-making basis for regional water resource optimization scheduling, water supply security, and refined management. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating the universal optimization method for water supply scheduling of parallel reservoir groups provided by the present invention. Figure 2 This is a schematic diagram of the parallel reservoir group water supply topology network provided by the present invention; Figure 3 This is a schematic diagram of the water storage results of the parallel reservoir group provided by the present invention; Figure 4 This is a schematic diagram of the water discharge results of the parallel reservoir group provided by the present invention; Figure 5 This is a schematic diagram of the water supply results of the parallel reservoir group provided by the present invention; Figure 6 This is a schematic diagram of the structure of the universal optimization system for water supply scheduling of parallel reservoir groups provided by the present invention; Figure 7 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0014] To address the shortcomings of existing parallel reservoir group water supply scheduling technologies, such as insufficient model topology representation capabilities, poor versatility leading to high customized development costs, and inconvenient switching of optimization algorithms, this invention proposes a universal optimization solution method for parallel reservoir group water supply scheduling.

[0015] Figure 1 This is a flowchart illustrating the general optimization method for water supply scheduling of parallel reservoir groups provided in this embodiment of the invention, as shown below. Figure 1 As shown, it includes: Step 100: Obtain the system operation data of each reservoir in the parallel reservoir group, and construct a generalized directed graph model based on the system operation data; Step 200: Based on the generalized generalized directed graph model, generalize the dynamic operation state of the parallel reservoir group water resource network, determine the directed edge flow matrix and node storage matrix, and deduce the dynamic relationship between the matrix elements through topological sorting and generalized water balance equation. Step 300: Based on the actual needs of water supply scheduling of the parallel reservoir group, dynamically select and parameterize the objective function and constraints, identify the scheduling decision variable matrix, quantify the degree of violation of each constraint, and form a general optimization problem definition; Step 400: Construct a parallel reservoir group scheduling decision evaluation module, receive the scheduling decision variable matrix, output the network dynamic operation state set through physical simulation calculation, and perform iterative solution based on the optimization algorithm module to obtain the optimal flow allocation scheme within the optimization period; Step 500: Select a representative solution using TOPSIS, take the system state driven by the optimal solution of the previous optimization cycle as the initial condition of the current optimization cycle, perform multi-cycle rolling optimization solution until the entire simulation period is completed, record the simulation results during the entire simulation period, and generate a detailed operation report.

[0016] Based on the above embodiments, step 100 includes: Acquire various basic data required for water supply scheduling of parallel reservoir groups, and construct a general topology network that can accurately reflect the physical connection and control relationship of the water resource system.

[0017] First, various system operation data of the study area are obtained. The data comes from structured documents such as historical watershed observation data, hydrological models, engineering design documents or socio-economic statistics.

[0018] Specifically, the embodiments of the present invention acquire engineering characteristic parameters of each reservoir in the study area, including the reservoir capacity corresponding to the normal water level, the reservoir capacity corresponding to the dead water level, the initial reservoir capacity during the optimization period, the maximum discharge capacity, and the minimum discharge requirement; acquire monthly water demand time series data and water consumption rate of each water user in the study area during the study period; and acquire the maximum flow limit of each water conveyance link in the study area and the monthly inflow and loss time series data during the study period.

[0019] Specifically, the research period for this embodiment is from 1956 to 2021.

[0020] Secondly, based on the acquired data, the physical entities in the parallel reservoir group are abstracted as nodes, and the hydraulic connections are abstracted as directed edges connecting the nodes, thereby constructing a topological network in the form of a multi-attribute heterogeneous directed graph.

[0021] This invention, for the first time in the field of water supply scheduling for parallel reservoir groups, proposes and constructs a multi-level, multi-attribute generalized directed graph model to generalize the complex water resource system topology and dynamic operating characteristics of parallel reservoir groups. The model can flexibly and accurately represent arbitrarily complex topologies such as multiple water sources, multiple water demand points, confluence and diversion, flow paths, and control relationships, thereby achieving model universality. Figure 2 As shown, the generalized directed graph model consists of three parts, and is generalized using the following mathematical expression: (1) in, For a set of nodes, It is an adjacency matrix. A multi-dimensional set of configurable attributes for nodes and directed edges.

[0022] Specifically, node set This is used to generalize the various physical entities in a water resource system and classify them into three node types based on node type. The node set... Generalize using the following mathematical expression: (2) in, It serves as a node in the reservoir, possessing the capacity for water storage and discharge. As a water-using node, it has water demand and return flow characteristics; These are confluence and diversion nodes, representing river merging, diversion, or canal system water distribution hubs; , , , , , These represent the total number of reservoir nodes, water user nodes, and diversion / merging nodes in the system, respectively.

[0023] Adjacency matrix This is used to generalize the hydraulic connections between all nodes in a water resources system and abstract them as directed edges connecting the nodes. Based on the set of nodes... The defined node order constructs the connection relationship into a... adjacency matrix ,in The total number of nodes in the system; if there are slave nodes in the system. To the node The direct hydraulic connection, then the matrix elements ,otherwise .

[0024] This is a multi-dimensional, configurable set of attributes related to nodes and directed edges, used to generalize the common parameters of nodes and directed edges in a water resources system. Based on the acquired data, the node set... Adjacency Matrix The parameters of the defined nodes and directed edges are matrixed or vectorized to construct the attribute sets of the nodes and directed edges in the model. The set of attributes Generalize using the following mathematical expression: (3) in, for Two-dimensional reservoir parameter matrix, Indicates the number of nodes in the reservoir. The number of reservoir parameters, matrix elements Used to store reservoir nodes The In this embodiment, the parameter values ​​are set for the reservoir node, and based on the acquired data, the reservoir capacity corresponding to the normal water level is determined. Dead water level corresponds to reservoir capacity Maximum discharge capacity Minimum discharge requirements Four parameters, namely ; for Two-dimensional water object parameter matrix, Indicates the number of nodes that use water. To represent the number of parameters for water-using objects, matrix elements Used to store water-using object nodes The In this embodiment, the parameter value is set for the water-using object node, based on the acquired data, to determine the water loss rate. One parameter, namely ; for The three-dimensional directed edge parameter matrix, This indicates the number of nodes in the water resource system. The number of parameters associated with directed edges, matrix elements Used to store directed edges The In this embodiment, the parameter values ​​are set for directed edges based on the acquired data. The maximum flow limit is a parameter, namely ; for A two-dimensional reservoir node inflow time series data matrix, The total length of the time series data, matrix elements Storage reservoir node During the period Inflow; for A two-dimensional water demand time series data matrix for water-using object nodes, matrix elements Storage water object node During the period Water demand; for A three-dimensional directed edge interval inflow time series data matrix, matrix elements Storing elements in the adjacency matrix The corresponding directed edge in the time period The inflow rate within the interval.

[0025] Based on the above multi-attribute directed graph generalization model By making full use of the acquired data, we can achieve unified, precise and configurable modeling of any complex topology and its operating characteristics. This not only enables efficient simulation of the hydraulic connections of complex parallel reservoir groups, but also forms a general parameterized expression framework for water resources systems, providing a solid and highly flexible general modeling foundation for solving the optimal scheduling of parallel reservoir groups.

[0026] Based on the above embodiments, step 200 includes: Based on the generalized directed graph model of the parallel reservoir group established in step 100, a directed edge flow matrix is ​​defined. and node water storage matrix This allows for the generalization of the dynamic operation of the parallel reservoir group water resource network and the description of the water volume changes of the entire water resource system over time.

[0027] Specifically, the directed edge flow matrix for A three-dimensional matrix, matrix elements Storage slave node To the node During the period Flow rate; node water storage matrix for A two-dimensional matrix, matrix elements storage nodes During the period The water storage capacity, and only when the node When the node is a reservoir, the matrix elements Greater than or equal to 0, otherwise Set to 0 to indicate that only the reservoir node has water storage capacity.

[0028] Based on water balance and combined with a directed graph generalization model Node water storage matrix Based on the directed edge flow matrix This determines the complete dynamic state of the entire network under a given directed edge flow. For any node in the generalized model... During the period Its water balance equation can be generalized as follows: (4) (5) (6) (7) in, For nodes During the period The water storage capacity, when hour, This is the initial state of the system, determined by the collected data; For nodes During the period The total traffic entering the node, For nodes During the period The total outflow of traffic from the nodes, For nodes During the period The total traffic consumed by the nodes.

[0029] Secondly, to ensure the accuracy of water balance calculations, a topological sorting method based on the Kahn algorithm is used to generalize the directed graph model. All nodes Based on the principle that the in-degree is 0, determine the time period. The order of water balance calculations. First, the generalized model... All nodes Calculate the in-degree and extract all nodes with an in-degree of 0. Add the nodes to the computation sequence; then, add the nodes. All outgoing adjacent nodes (Right now corresponding nodes Decrement the in-degree of 1 if any adjacent node at this time If the in-degree of a node becomes 0, it is added to the computation sequence; this step is repeated until all nodes are included. The calculation order is determined for each node. This order ensures that, in the water balance calculation, the water volume calculation for each node is based on the determined outflows of all its upstream nodes. Specifically, this is based on the adjacency matrix. Calculate the in-degree of each node: (8) Through the above steps, in the directed graph generalization model Based on this, a generalized expression of the dynamic operation status of water supply scheduling of parallel reservoir groups was realized; by topological sorting and general water balance equations, the relationship between matrix elements was calculated, so that the complex dynamic behavior of water resources system can be described in a structured and mathematical way, providing a unified format for subsequent optimization solutions, thereby realizing generalized optimization scheduling.

[0030] Based on the above embodiments, step 300 includes: The actual water supply scheduling needs of the parallel reservoir group are clearly defined, and an optimization objective function is dynamically selected and parameterized. The optimization objective function adopts a modular design and is implemented through a unified interface, allowing users to flexibly select and extend new objective functions according to specific needs. The set of optimization objective functions can be summarized as follows: (9) in, This can include minimizing the total water shortage for water users, minimizing the amount of water released from reservoirs, etc. Each objective function... It can be represented as the directed edge flow matrix defined in step 200. and node water storage matrix The functional relationship between them is determined, and the multi-objective function values ​​are further obtained. This constitutes a complete scheduling objective function module.

[0031] In this embodiment of the invention, the optimization objective function includes: Minimize the total water shortage, which means minimizing the total unmet demand of all water users during the optimization period. Its mathematical expression is: (10) Minimize the total water discharge, specifically the total water discharge from all reservoirs during the optimization period. Its mathematical expression is: (11) in, For reservoir j During the period t The amount of water discarded is determined based on the directed graph generalization model. The reservoir node parameter matrix and water storage at each node during different time periods Sure.

[0032] To maximize the reservoir's final water storage volume during the storage period. Its mathematical expression is: (12) in, For reservoir j At the end of its water storage period The water storage capacity.

[0033] Based on the actual needs of water supply scheduling for the parallel reservoir group, one or more constraints are dynamically selected, and the selected constraints are parameterized. The constraints also adopt a modular design, implemented through a unified interface, supporting users to dynamically select and extend new constraints. In this embodiment, the selected constraints include: Based on the engineering characteristic parameters of each reservoir in the study area, the normal water level corresponding to the reservoir's capacity and the dead water level corresponding to the reservoir's capacity are used to constrain the water storage volume of each reservoir at each time period, ensuring that the water storage volume of each reservoir does not exceed the maximum capacity and is not lower than the dead capacity at any time. The mathematical expression is: (13) Based on the obtained engineering characteristic parameters of each reservoir in the study area, the maximum and minimum discharge limits of each reservoir are used to constrain the outflow of water from each reservoir at each time period, ensuring that the outflow of water from each reservoir does not exceed the maximum discharge capacity and is not lower than the minimum discharge requirement at any time. The mathematical expression is: (14) Based on the engineering characteristic parameters of each water conveyance link in the study area, the maximum flow limit of directed edges is used to constrain the water conveyance volume of each directed edge at each time period, ensuring that the flow of the water conveyance link does not exceed its maximum transmission capacity at any time period. Its mathematical expression is: (15) Based on the constraints, define the violation quantity. This is used to quantify the degree of violation of each constraint. The mathematical expression for the violation is as follows: (16) in, V The total penalty value. Constraints c The penalty factor Constraints c The amount of violation, When the corresponding physical quantity violates the constraint, the difference between the corresponding physical quantity and the corresponding extreme value is taken; when the constraint is not violated, the value is 0. C} represents the set of constraints.

[0034] In this embodiment, the mathematical expressions for the violation values ​​of each selected constraint are as follows: (17) (18) (19) Secondly, based on the directed graph generalization model constructed in step 100 and the generalized directed edge flow matrix in step 200 Based on the proactive decision-making principle in water resources systems, the unit-time scale flow of directed edges related to reservoir outflow and water intake by water users is defined as the scheduling decision variable. The set of scheduling decision variables can be represented as a decision variable matrix. ,in For directed edges identified as decision variables in time period The traffic. Specifically, based on the adjacency matrix. The definition of decision variables includes the following rules: First, for the adjacency matrix... elements in When node That is, the node When it is a reservoir node, define The corresponding directed edge in any time period Traffic flow is the decision variable To characterize the reservoir's proactive water release decision; if the node , i.e., node Define a water-using object node. The corresponding directed edge in any time period Traffic flow is the decision variable This is used to characterize the proactive water intake decisions of water users. The final decision variable matrix is ​​then determined. The decision variable matrix is ​​determined through optimization algorithms. The optimization result will be the directed edge flow matrix. The flow values ​​of each corresponding directed edge in each time period are replaced with This allows updating the directed graph generalization model. Water storage matrix at each time period This allows us to determine the objective function value and constraint violation level at the current optimization result. Furthermore, the flow rates of the directed edges corresponding to non-decision variables at each time period are derived from the above process using formula (4) based on water balance.

[0035] Figure 3 , Figure 4 and Figure 5 The heatmaps show a comparison of the water storage capacity of all reservoirs in the parallel reservoir group water supply topology network, a comparison of the water discharge capacity of all reservoirs, and a comparison of the water supply satisfaction rate of all water users.

[0036] Based on the above embodiments, step 400 includes: The first step is to generalize the directed graph model constructed in step 100. and the directed edge flow matrix generalized in step 200 and node water storage matrix The components are integrated to form a physical simulation module for water supply scheduling of a parallel reservoir group. This module can receive the decision variable matrix generated by the optimization algorithm, and thus update the directed edge flow matrix according to step 200. and node water storage matrix Simulation of a generalized directed graph model Status at different times.

[0037] The second step involves integrating the optimization objective function and constraints defined in step 300 to form a parallel reservoir group scheduling decision evaluation module. This evaluation module receives scheduling decision variables generated by the optimization algorithm, calculates the values ​​of each optimization objective function based on simulation results, and generates corresponding violation values ​​according to constraint violations.

[0038] The third step involves constructing an optimization algorithm set module for a given optimization cycle, selecting the appropriate module based on the optimization requirements. This optimization algorithm module can be either a multi-objective optimization algorithm or a single-objective optimization algorithm. For multi-objective optimization algorithms, this embodiment selects the NSGA-III algorithm. This algorithm directly uses the multiple optimization objective functions configured in step 300 as optimization objectives, and iteratively searches for the Pareto optimal solution set in the multi-objective search space through its internal non-dominated sorting and crowding distance calculation mechanisms. For single-objective optimization algorithms, this embodiment selects particle swarm optimization, simulated annealing, and dynamic programming algorithms to form the algorithm set module. This module integrates the multiple optimization objective functions configured in step 300 into a single comprehensive objective function through weighted summation, where the sum of the weights of each objective function is 1.

[0039] During the iterative solution process, the optimization algorithm module integrates the constructed scheduling decision evaluation module. Specifically, for search-based optimization algorithms, each time a set of scheduling decision variables is generated, it is input into the scheduling decision evaluation module. The evaluation module calculates and outputs the corresponding objective function value (or comprehensive objective function value) and constraint violation penalty term by calling the water resources physical simulation module. For dynamic programming algorithms, during the stage-by-stage and state-by-state solution process, the objective function and constraint condition logic in the scheduling decision evaluation module are called to calculate the state transition cost of each decision. Based on the received objective function value and penalty term (or cost), the optimization algorithm module adjusts the scheduling decision variables and repeats the above evaluation process until the algorithm's own convergence condition is met or the preset maximum number of iterations is reached, thereby obtaining the optimal flow allocation scheme for each decision variable in the water conveyance link within the optimization cycle.

[0040] Based on the above embodiments, step 500 includes: The entire simulation period is divided into multiple consecutive optimization cycles. In this embodiment of the invention, a single optimization cycle is selected to be 12 months in length, with an optimization scale of monthly. The water supply scheduling problem of the parallel reservoir group in the study area is optimized and solved during the study period. If it is the first optimization cycle, the initial conditions input in step 100 are used as the initial system state; if it is a subsequent optimization cycle, the parallel reservoir group system state driven by the optimal solution at the end of the previous optimization cycle is used as the initial condition of the current optimization cycle.

[0041] Based on the solution method for a single optimization cycle described in step 400, the solution is applied to each optimization cycle until the entire simulation period is completed. After each optimization cycle is solved, if the optimization algorithm is a multi-objective optimization algorithm, it will generate a Pareto optimal solution set. In this embodiment of the invention, selection is based on the ideal solution similarity (TOPSIS). This method calculates the distance between each Pareto solution and the ideal positive solution (PIS) and the ideal negative solution (NIS), and selects the solution that is closest to the PIS and furthest from the NIS as the representative solution. Specifically, for the Pareto front... In One solution And each solution has The optimization objective function value The steps for selecting TOPSIS are as follows: The first step is to normalize each objective function value to eliminate the influence of dimensions and construct a normalized decision matrix. Its mathematical expression is: (20) The second step is to assign weights to each objective based on its importance. Construct a weighted normalized decision matrix weights and Its mathematical expression is: (twenty one) The third step is to determine the ideal correct solution. and ideal negative solution For each target , the ideal solution The optimal value of all solutions on the objective is the ideal negative solution. The worst value is expressed mathematically as follows: (twenty two) The fourth step is to calculate the distance from each solution to the ideal positive and ideal negative solutions, which is expressed mathematically as follows: (twenty three) Fifth step: Calculate the relative proximity of each solution. Its mathematical expression is: (twenty four) Final choice The solution with the largest value is taken as the representative solution.

[0042] After each optimization cycle is completed, detailed simulation results for that cycle are recorded, including the water storage of each reservoir, the supply and demand balance of each water user, the flow rate of each water conveyance link, and any violations of constraints. Upon completion of the entire simulation period, the simulation results from all optimization cycles are summarized and compiled, and a detailed operation report is generated. This report includes comprehensive performance indicators such as total reservoir water supply, total water wastage, total water shortage for each water user, and water supply guarantee rate, to support decision analysis and scheme comparison.

[0043] Overall, this embodiment provides a general optimization solution method for water supply scheduling of parallel reservoir groups. The specific solution process is as follows: First, obtain various system operation data required for scheduling of parallel reservoir groups, and based on the obtained data, construct a multi-level, multi-attribute generalized directed graph model. This enables unified, precise, and configurable modeling of arbitrarily complex topologies and their operational characteristics. Subsequently, based on the generalized model, the dynamic operational state of the water resource network is defined and generalized, including the directed edge flow matrix. and node water storage matrix The system determines the node computation order through topological sorting and calculates the dynamic relationships between matrix elements using a general water balance equation, thereby describing the water volume changes of the entire water resource system over time. Based on actual scheduling needs, it dynamically selects and parameterizes the objective function and constraints for optimization, and identifies and defines the scheduling decision variable matrix. Furthermore, the degree of violation of each constraint is quantified to form a generalized optimization problem definition; furthermore, a parallel reservoir group scheduling decision evaluation module is constructed, which generalizes the directed graph model. As static input, it receives the decision variable matrix generated by the optimization algorithm. By calling the integrated water resources physical simulation module, based on topological sorting and the general water balance equation, the dynamic operating state set of the network is calculated and updated, and the optimization objective function value set and the total constraint violation penalty value are calculated based on this. For an optimization cycle, an appropriate optimization algorithm module (including multi-objective evolutionary algorithm or single-objective optimization algorithm) is selected and iteratively solved in combination with the constructed scheduling decision evaluation module to obtain the optimal flow allocation scheme for each decision variable in the optimization cycle. Among them, the multi-objective optimization algorithm (such as NSGA-III) will generate a Pareto optimal solution set, and the representative solution is selected by TOPSIS. Finally, the entire simulation period is divided into multiple consecutive optimization cycles, and multi-cycle rolling optimization is performed. The system state driven by the optimal scheme of the previous cycle is used as the initial condition of the current cycle until the entire simulation period is completed. Detailed simulation results in all optimization cycles are recorded, and a detailed operation report including comprehensive performance indicators such as total reservoir water supply, total water abandonment, total water shortage of water users, and water supply guarantee rate is generated.

[0044] As can be seen, the general optimization solution method for water supply scheduling of parallel reservoir groups provided in this embodiment achieves effective decoupling and unified integration of data input, topology construction, optimization problem definition, physical simulation, and optimization solution through modular design. This method can efficiently and accurately construct and simulate the complex topological structures of parallel reservoir groups with multiple water sources, multiple water demand points, and confluence / diversion flows. It also supports dynamic configuration of optimization objectives and constraints, and flexible integration of multiple optimization algorithms. The method integrates the water resources physical simulation module into the evaluation stage of the optimization algorithm, ensuring accurate evaluation of scheduling decisions in the physical process. Through a multi-period rolling optimization strategy, it obtains a more adaptable and robust scheduling scheme. This invention effectively solves the shortcomings of existing technologies, such as poor model versatility, high customized development costs, and inconvenient switching of optimization algorithms, providing important technical support for regional water resources optimization and refined management.

[0045] The following describes the universal optimization system for water supply scheduling of parallel reservoir groups provided by the present invention. The universal optimization system for water supply scheduling of parallel reservoir groups described below can be referred to in correspondence with the universal optimization method for water supply scheduling of parallel reservoir groups described above.

[0046] Figure 6 This is a schematic diagram of the structure of the universal optimization system for water supply scheduling of parallel reservoir groups provided in an embodiment of the present invention, as shown below. Figure 6 As shown, it includes: a data input module 61, a topology construction module 62, a problem evaluation module 63, an optimization solution module 64, and a result reporting module 65, wherein: The data input module 61 is used to acquire system operation data of each reservoir in the parallel reservoir group, and construct a generalized directed graph model based on the system operation data; the topology construction module 62 is used to generalize the dynamic operation state of the water resource network of the parallel reservoir group based on the generalized directed graph model, determine the directed edge flow matrix and node storage matrix, and deduce the dynamic relationship between the matrix elements through topology sorting and generalized water balance equation; the problem evaluation module 63 is used to dynamically select and parameterize the optimization objective function and constraints according to the actual needs of water supply scheduling of the parallel reservoir group, identify the scheduling decision variable matrix, and quantify the violation of each constraint. The degree of optimization is used to form a generalized definition of the optimization problem; the optimization solution module 64 is used to construct the parallel reservoir group scheduling decision evaluation module, receive the scheduling decision variable matrix, output the network dynamic operation state set through physical simulation calculation, and perform iterative solution based on the optimization algorithm module to obtain the optimal flow allocation scheme within the optimization period; the result reporting module 65 is used to select representative solutions through TOPSIS, take the system state driven by the optimal scheme of the previous optimization period as the initial condition of the current optimization period, perform multi-period rolling optimization solution until the entire simulation period is completed, record the simulation results within the entire simulation period, and generate a detailed operation report.

[0047] Figure 7 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 7 As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730, and a communication bus 740. The processor 710, communication interface 720, and memory 730 communicate with each other via the communication bus 740. The processor 710 can call logical instructions in the memory 730 to execute a general optimization method for water supply scheduling of a parallel reservoir group. This method includes: acquiring system operation data of each reservoir in the parallel reservoir group; constructing a generalized directed graph model based on the system operation data; generalizing the dynamic operating state of the water resource network of the parallel reservoir group based on the generalized directed graph model; determining the directed edge flow matrix and node storage matrix; and calculating the dynamic relationships between matrix elements through topological sorting and a generalized water balance equation; dynamically selecting and parameterizing the optimization objective function and constraints according to the actual needs of water supply scheduling of the parallel reservoir group; and identifying... A scheduling decision variable matrix is ​​constructed, and the degree of violation of each constraint is quantified to form a general optimization problem definition. A parallel reservoir group scheduling decision evaluation module is built, which receives the scheduling decision variable matrix, outputs the dynamic operating state set of the network through physical simulation calculation, and performs iterative solution based on the optimization algorithm module to obtain the optimal flow allocation scheme within the optimization period. A representative solution is selected through TOPSIS, and the system state driven by the optimal scheme of the previous optimization period is used as the initial condition of the current optimization period. Multi-period rolling optimization is performed until the entire simulation period is completed, and the simulation results within the entire simulation period are recorded to generate a detailed operation report.

[0048] Furthermore, the logical instructions in the aforementioned memory 730 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0049] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0050] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0051] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A general optimization method for water supply scheduling of parallel reservoir groups, characterized in that, include: Obtain system operation data of each reservoir in a parallel reservoir group, and construct a generalized directed graph model based on the system operation data; Based on the generalized generalized directed graph model, the dynamic operation state of the parallel reservoir group water resource network is generalized, the directed edge flow matrix and node storage matrix are determined, and the dynamic relationship between the matrix elements is deduced through topological sorting and general water balance equation. Based on the actual needs of water supply scheduling of parallel reservoir groups, the objective function and constraints are dynamically selected and parameterized for optimization, the scheduling decision variable matrix is ​​identified, and the degree of violation of each constraint is quantified to form a general optimization problem definition. A parallel reservoir group scheduling decision evaluation module is constructed, which receives the scheduling decision variable matrix, outputs the dynamic operation state set of the network through physical simulation calculation, and performs iterative solution based on the optimization algorithm module to obtain the optimal flow allocation scheme within the optimization period. The TOPSIS algorithm, which approximates ideal solutions, selects representative solutions. The system state driven by the optimal solution in the previous optimization cycle is used as the initial condition for the current optimization cycle. Multi-cycle rolling optimization is performed until the entire simulation period is completed. The simulation results during the entire simulation period are recorded, and a detailed operation report is generated.

2. The universal optimization method for water supply scheduling of parallel reservoir groups according to claim 1, characterized in that, Acquire system operation data of each reservoir in a parallel reservoir group, and construct a generalized directed graph model based on the system operation data, including: The system operation data includes engineering characteristic parameters, water user demand time series, water transmission link transmission capacity, and water input and loss. The system operation data comes from a preset structured file. Based on the system operation data, the physical entities in the parallel reservoir group are abstracted as nodes, and the hydraulic connections are abstracted as directed edges connecting the nodes, thus forming a generalized directed graph model in the form of a multi-attribute heterogeneous directed graph: in, For a set of nodes, It is an adjacency matrix. A multi-dimensional set of configurable attributes for nodes and directed edges; Node set This is used to generalize the various physical entities in a water resource system, and classifies them into three node types based on node type, using the following mathematical expression for generalization: in, It serves as a node in the reservoir, possessing the capacity for water storage and discharge. As a water-using node, it has water demand and return flow characteristics; These are confluence and diversion nodes, representing river merging, diversion, or canal system water distribution hubs; , , , , , These represent the total number of reservoir nodes, water user nodes, and diversion / merging nodes in the system, respectively. Adjacency matrix This is used to generalize the hydraulic connections of all nodes in a water resources system and abstract them as directed edges connecting the nodes, constructing the connection relationships as a... adjacency matrix ,in The total number of nodes in the system; if there are slave nodes in the system. To the node The direct hydraulic connection, then the matrix elements ,otherwise ; The set of multi-dimensional, configurable attributes associated with nodes and directed edges can be generalized using the following mathematical expression: in, for Two-dimensional reservoir parameter matrix, Indicates the number of nodes in the reservoir. The number of reservoir parameters, matrix elements Used to store reservoir nodes The Each parameter value; for Two-dimensional water object parameter matrix, Indicates the number of nodes that use water. To represent the number of parameters for water-using objects, matrix elements Used to store water-using object nodes The Each parameter value; for The three-dimensional directed edge parameter matrix, This indicates the number of nodes in the water resource system. The number of parameters associated with directed edges, matrix elements Used to store directed edges The Each parameter value; for A two-dimensional reservoir node inflow time series data matrix, The total length of the time series data, matrix elements Storage reservoir node During the period Inflow; for A two-dimensional water demand time series data matrix for water-using object nodes, matrix elements Storage water object node During the period Water demand; for A three-dimensional directed edge interval inflow time series data matrix, matrix elements Storing elements in the adjacency matrix The corresponding directed edge in the time period The inflow rate within the interval.

3. The universal optimization method for water supply scheduling of parallel reservoir groups according to claim 1, characterized in that, Based on the generalized directed graph model, the dynamic operating state of the parallel reservoir group water resource network is generalized, the directed edge flow matrix and node storage matrix are determined, and the dynamic relationships between matrix elements are calculated through topological sorting and the general water balance equation, including: Define the directed edge flow matrix and node water storage matrix Generalize the dynamic operating state of the water resource network of a parallel reservoir group; directed edge flow matrix for A three-dimensional matrix, matrix elements Storage slave node To the node During the period Flow rate; node water storage matrix for A two-dimensional matrix, matrix elements storage nodes During the period The water storage capacity, and only when the node When the node is a reservoir, the matrix elements Greater than or equal to 0, otherwise Set to 0 to indicate that only the reservoir node has water storage capacity; Based on water balance and combined with a generalized directed graph model The directed edge flow matrix Determine the node water storage matrix It expresses the dynamic operating state of the entire network under a given directed edge flow; for any node in the generalized directed graph model During the period The corresponding water balance equation can be simplified as follows: in, For nodes During the period The water storage capacity, when hour, This is the initial state of the system, determined by the collected data; For nodes During the period The total traffic entering the node, For nodes During the period The total outflow of traffic from the nodes, For nodes During the period The total traffic consumed by the nodes; To ensure water balance, a topological sorting method based on Kahn's algorithm is used to generalize the directed graph model. All nodes Based on the principle that the in-degree is 0, determine the time period. The order of water balance calculations; first, the generalized model. All nodes Calculate the in-degree and extract all nodes with an in-degree of 0. Add the nodes to the computation sequence; then, add the nodes. All outgoing adjacent nodes Decrement the in-degree by 1 if any adjacent node at this time If the in-degree of a node becomes 0, it is added to the computation sequence; this step is repeated until all nodes are included. The calculation order is determined for all cases; among them, the adjacency matrix is ​​used to determine the order of calculation. Calculate the in-degree of each node. : 。 4. The universal optimization method for water supply scheduling of parallel reservoir groups according to claim 1, characterized in that, Based on the actual needs of water supply scheduling in a parallel reservoir group, the objective function and constraints are dynamically selected and parameterized for optimization. The scheduling decision variable matrix is ​​identified, and the degree of violation of each constraint is quantified, forming a generalized optimization problem definition, including: The system dynamically selects and parameterizes the selected objective function, employing a modular design and a unified interface. This allows users to flexibly select and extend new objective functions according to specific needs. The set of objective functions is summarized as follows: in, Each objective function includes minimizing the total water shortage of water users and minimizing the water release from reservoirs. Represented as a directed edge flow matrix and node water storage matrix The functional relationship between them yields the multi-objective function values. This constitutes a complete scheduling objective function module; Based on the actual needs of water supply scheduling in a parallel reservoir group, one or more constraints are dynamically selected, and the selected constraints are parameterized. A modular design is adopted, implemented through a unified interface, supporting users to dynamically select and extend new constraints; violation quantities are defined. This is used to quantify the degree of violation of each constraint; the violation quantity The mathematical expression is as follows: in, V The total penalty value. Constraints c The penalty factor Constraints c The amount of violation, When the corresponding physical quantity violates the constraint, the difference between the corresponding physical quantity and the corresponding extreme value is taken; when the constraint is not violated, the value is 0. C } represents the set of constraints; Based on the generalization model of directed graphs and directed edge flow matrix Based on the proactive decision-making principle in water resources systems, the unit-time scale flow of directed edges related to reservoir outflow and water intake by water users is defined as the scheduling decision variable, and is represented as a decision variable matrix. ,in For directed edges identified as decision variables in time period Traffic.

5. The universal optimization method for water supply scheduling of parallel reservoir groups according to claim 1, characterized in that, A parallel reservoir group scheduling decision evaluation module is constructed. This module receives the scheduling decision variable matrix, outputs a dynamic network operating state set through physical simulation calculations, and iteratively solves the problem based on an optimization algorithm module to obtain the optimal flow allocation scheme within the optimization period. This includes: The generalized model of the constructed directed graph Directed edge flow matrix and node water storage matrix The system is integrated to form a physical simulation module for water supply scheduling of a parallel reservoir group. This module receives the decision variable matrix generated by the optimization algorithm and, based on the directed edge flow matrix... and node water storage matrix Simulation of a generalized directed graph model The status at each time period; The optimization objective function and constraints are integrated to form a parallel reservoir group scheduling decision evaluation module. The parallel reservoir group scheduling decision evaluation module receives the scheduling decision variables generated by the optimization algorithm, calculates the values ​​of each optimization objective function based on the simulation results, and generates corresponding violation values ​​according to the constraint violation situation. For an optimization cycle, an optimization algorithm set module is constructed, and the optimization algorithm module is selected according to the optimization requirements. The optimization algorithm set module can be any one of a multi-objective optimization algorithm or a single-objective optimization algorithm. During the iterative solution process, the constructed parallel reservoir group scheduling decision evaluation module is integrated into the optimization algorithm set module. Based on the received objective function value and penalty term, the scheduling decision variables are adjusted, and the above evaluation process is repeated until the convergence condition of the algorithm itself is met or the preset maximum number of iterations is reached, so as to obtain the optimal flow allocation scheme within the optimization cycle.

6. The universal optimization method for water supply scheduling of parallel reservoir groups according to claim 1, characterized in that, Representative solutions are selected using TOPSIS. The system state driven by the optimal solution in the previous optimization cycle is used as the initial condition for the current optimization cycle. Multi-cycle rolling optimization is performed until the entire simulation period is completed. The simulation results throughout the entire simulation period are recorded, and a detailed operation report is generated, including: The entire simulation period is divided into multiple consecutive optimization cycles, and the water supply scheduling problem of the parallel reservoir group in the study area is optimized and solved within the study period. If it is the first optimization cycle, the input initial conditions are used as the initial system state. Based on the described solution method for a single optimization cycle, each optimization cycle is solved until the entire simulation period is completed. After the solution of each optimization cycle is completed, if the optimization algorithm is a multi-objective optimization algorithm, the optimal solution is selected based on the TOPSIS method. After each optimization cycle is completed, detailed simulation results are recorded, including the water storage of each reservoir, the supply and demand balance of each water user, the flow rate of each water conveyance link, and the violation of all constraints. After the entire simulation period is completed, the simulation results of all optimization cycles are summarized and organized, and a detailed operation report is generated. The operation report includes comprehensive performance indicators such as the total water supply of the reservoir, the total water wastage, the total water shortage of the water user, and the water supply guarantee rate, to support decision analysis and scheme comparison.

7. A universal optimization system for water supply scheduling of parallel reservoir groups, characterized in that, include: The data input module is used to acquire system operation data of each reservoir in the parallel reservoir group and to construct a generalized directed graph model based on the system operation data. The topology construction module is used to generalize the dynamic operation state of the parallel reservoir group water resources network based on the generalized generalized model of the generalized directed graph, determine the directed edge flow matrix and the node water storage matrix, and deduce the dynamic relationship between the matrix elements through topological sorting and generalized water balance equation. The problem assessment module is used to dynamically select and parameterize the objective function and constraints for optimization based on the actual needs of water supply scheduling of the parallel reservoir group, identify the scheduling decision variable matrix, quantify the degree of violation of each constraint, and form a general optimization problem definition. The optimization solution module is used to construct the parallel reservoir group scheduling decision evaluation module. It receives the scheduling decision variable matrix, outputs the network dynamic operation state set through physical simulation calculation, and performs iterative solution based on the optimization algorithm module to obtain the optimal flow allocation scheme within the optimization period. The results reporting module is used to select representative solutions through TOPSIS, take the system state driven by the optimal solution of the previous optimization cycle as the initial condition of the current optimization cycle, perform multi-cycle rolling optimization solutions until the entire simulation period is completed, record the simulation results during the entire simulation period, and generate a detailed operation report.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the general optimization method for water supply scheduling of parallel reservoir groups as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the general optimization method for water supply scheduling of parallel reservoir groups as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the general optimization method for water supply scheduling of parallel reservoir groups as described in any one of claims 1 to 6.

Images