Multi-constraint particle swarm optimization electric power overhaul scheduling method, equipment and medium

Abstract

The invention provides a multi-constraint particle swarm optimization electric power overhaul scheduling method and device and a medium, and the method combines a dynamic inertia weight adjustment mechanism and a midpoint guide variation strategy to enhance the global search capability. And a genetic evolution mechanism of elite retention, random mask crossover and adaptive probability variation is introduced to improve population diversity and optimization stability. In addition, a multi-neighborhood local hill-climbing search and multi-constraint feasibility repair function is designed, and it is ensured that the scheduling scheme has high quality on the premise that constraints are met. Experimental results show that the method is superior to a traditional algorithm in the aspects of convergence speed, solution feasibility and scheduling quality, and has good engineering applicability and optimization effect.

Description

Technical Field

[0001] This research relates to the field of distribution network maintenance planning, specifically to a multi-constraint particle swarm optimization algorithm for distribution network maintenance scheduling that integrates genetic inspiration and dynamic penalty mechanisms, as well as its model optimization strategy. Background Technology

[0002] As a crucial component of the power system, the distribution network undertakes the critical tasks of power transmission and redistribution, serving as a vital bridge connecting power sources and users. With the continuous expansion of my country's power grid and the rapid growth in the number of power equipment, the distribution network structure is becoming increasingly complex. Under sudden events such as natural disasters and extreme weather, distribution network equipment is prone to failure, necessitating power outages for maintenance of relevant lines and equipment to ensure system safety. The rationality of the maintenance plan directly affects the scope and duration of power outages, and has a profound impact on the safety, economy, and reliability of the power system. Therefore, scientifically formulating maintenance plans not only helps ensure the safe and stable operation of the power grid but also helps reduce power outage losses for users and improve maintenance efficiency and economic benefits.

[0003] Optimizing distribution network maintenance plans is a multi-objective, multi-constraint nonlinear integer programming problem. It requires balancing multiple objectives, such as maintenance economy, operational risk, and scheduling rationality, while also satisfying constraints related to maintenance resources, maintenance relationships, and system security. In recent years, scholars both domestically and internationally have proposed various mathematical modeling and intelligent solution methods for distribution network maintenance optimization. For example, based on the multi-objective particle swarm optimization algorithm, optimization modeling was performed to address maintenance economy and power supply loss; from the perspective of operational risk, a maintenance optimization model with the objective of minimizing risk was constructed, and an improved optimization algorithm was designed for solving it; and considering load transfer characteristics, a two-level optimization method was used to optimize maintenance scheduling.

[0004] Particle Swarm Optimization (PSO) algorithms are widely used in power maintenance and scheduling optimization due to their advantages of few parameters, fast convergence speed, and ease of implementation. However, standard PSO is prone to getting trapped in local optima and insufficient search diversity when dealing with complex, multi-constraint maintenance and scheduling problems. To address these shortcomings, this paper proposes a multi-constraint PSO model that integrates genetic heuristics and dynamic penalty mechanisms. This method enhances global search capabilities by dynamically adjusting inertia weights, introduces a heuristic genetic evolution mechanism to improve population diversity, and incorporates multi-neighborhood local search and feasibility repair strategies in the update iterations to better balance maintenance economy and system safety. Finally, the proposed model and algorithm are validated using a real distribution network example. The results show that this method outperforms traditional methods in terms of optimization accuracy, convergence speed, and feasible solution quality, demonstrating high engineering application value. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of the prior art by proposing a multi-constraint particle swarm optimization power maintenance scheduling method. This method structurally improves the traditional particle swarm optimization algorithm by introducing dynamic weight adjustment, adaptive crossover mutation, multi-neighborhood search, and feasibility repair mechanisms, thereby significantly improving the solution quality and convergence performance of multi-objective, multi-constraint scheduling problems.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: This invention provides a multi-constraint particle swarm optimization method for power maintenance scheduling, which includes the following steps: Step 1: Construct a power distribution network maintenance scheduling optimization model. The optimization model aims to minimize power outage losses and maintenance costs. It considers five types of constraints in the actual maintenance process, namely simultaneous constraints, mutual exclusion constraints, sequential constraints, resource constraints, and time window constraints, forming a multi-objective nonlinear integer programming problem.

[0007] Step 2: Initialize and improve particle swarm optimization parameters, setting parameters such as the number of particles, maximum number of iterations, velocity limit, learning factor, and initial inertia weight. Encode particle positions as maintenance start time vectors, construct the particle swarm population, and initialize the fitness function.

[0008] Step 3: Optimize the distribution network maintenance model using an improved particle swarm optimization algorithm to avoid problems such as getting trapped in local optima and insufficient search diversity that may occur with traditional optimization algorithms, and obtain the optimized distribution network maintenance plan arrangement results.

[0009] Step 4: Introduce a heuristic genetic evolution mechanism into the improved particle swarm framework, and integrate elite retention strategy, heuristic initialization, adaptive crossover mutation and local hill climbing search to construct a hybrid global jump operator.

[0010] Step 5: After reaching the maximum number of iterations or the convergence condition, output the optimal scheduling particle, decode it into the corresponding equipment maintenance plan, and form an executable scheduling scheme that meets the constraints and optimization objectives.

[0011] Furthermore, the model structure in step 1 of the present invention is as follows: The objective function to be optimized is denoted as:

[0012] in, Power outage losses due to maintenance; Maintenance costs for each piece of equipment during the planned period.

[0013] Specifically, it is expressed as follows:

[0014] in, Total number of equipment to be inspected; : Maintenance start time of device i; Maintenance duration of device i (in time period); The economic loss caused by the load interruption due to the maintenance of equipment i in time period t; : The direct economic cost (such as labor, materials, tools, etc.) of equipment i performing maintenance in time period t; Simultaneous constraint: In the maintenance and scheduling of power systems, if two or more devices have control interlocking relationships, are geographically close, or belong to the same operating unit (such as the same line or the same bay), these devices should be maintained within the same time period to avoid repeated power outages, improve maintenance efficiency, and reduce the scope of power outages.

[0015] Therefore, suppose there exists a set of equipment pairs that require simultaneous maintenance:

[0016] Each pair Indicates equipment and Simultaneous maintenance is required. Therefore, the mathematical form of this constraint can be expressed as:

[0017] Mutual Exclusion Constraint: During power distribution network maintenance, some devices may share critical facilities such as protection devices and circuit breakers, as they support the same power supply path within the system structure. If these devices are maintained simultaneously, it may lead to a complete interruption of the power supply path or failure of the protection configuration. Therefore, devices with such logically mutually exclusive relationships must not be maintained simultaneously during overlapping time periods; that is, their maintenance time periods must not overlap at all.

[0018] Let the set of device pairs that need to satisfy mutual exclusion be:

[0019] in Indicates equipment With equipment There is a mutual exclusion relationship, and their maintenance start times are defined as follows: , The maintenance time is as follows: , Then the mathematical expression for the mutual exclusion constraint is:

[0020] Maintenance time window constraints: Each maintenance task must be completed within the specified time window to minimize the impact on the power system.

[0021]

[0022] in and These are the earliest and latest start times for the k-th maintenance item, respectively. This refers to the actual maintenance time.

[0023] Maintenance resource constraints: The manpower, material resources, equipment and financial resources required for maintenance are limited and need to be rationally allocated under the given resource conditions.

[0024]

[0025] in It is the r-th type of resource used in the maintenance plan. It is the total amount of this type of resource.

[0026] Sequence Constraints: In power distribution systems, the maintenance of some critical equipment has sequential dependencies. For example, busbar maintenance can only begin after the circuit breaker has completed its isolation operation, or certain interval maintenance can only be carried out after the main transformer has been completed. In such scenarios, to ensure maintenance safety and standardized work procedures, it is necessary to set sequence constraints for equipment maintenance. Assume there exists a set of equipment pairs that must satisfy the sequence relationship:

[0027] Each pair Indicates equipment Maintenance must precede equipment repair. Finish. :equipment The start time of maintenance; :equipment Maintenance time; :equipment The start time of maintenance.

[0028] Furthermore, the improved particle swarm optimization algorithm in step 3 of the present invention is as follows: Step 3-1: Dynamic Inertia Weight Adjustment Mechanism To achieve a balance between global and local search capabilities, an exponentially decreasing dynamic inertia factor is introduced, defined as follows:

[0029] in, Let be the inertia factor at the t-th iteration. and The maximum and minimum values ​​of the inertia factor are taken in this paper. =0.9, =0.2; The decay rate factor is set to 0.1 in this paper.

[0030] Step 3-2: Midpoint-guided mutation speed update strategy.

[0031] Based on the standard PSO velocity update formula, this paper introduces an individual optimal solution. With the global optimal solution The midpoint serves as the guiding center:

[0032] The formulas for updating the particle's velocity and position are:

[0033]

[0034] in, , is a uniformly distributed random number; and The learning factor (usually 2); This is the feasibility repair function. Since the position updates are continuous values, they need to be discretized in the scheduling problem to ensure that the constraints are met.

[0035] Step 3-3 Feasibility Repair Function repair() 1. Legality of maintenance period restored:

[0036] Ensure that the scheduling time does not exceed the limit.

[0037] 2. Simultaneous constraint repair: If devices i and j have simultaneous constraints:

[0038] Directly Assigned value To achieve synchronization.

[0039] 3. Mutual Exclusion Constraint Repair: If the two maintenance periods overlap:

[0040] The adjustment is as follows: Or postpone to the next feasible time. 4. Sequence constraint repair: To ensure that i completes before j:

[0041] If not satisfied, adjust as follows:

[0042] 5. Maximum number of concurrent devices: For example, the maximum number of devices that can be repaired in each time period is Then, the usage for each time period is counted, and if the limit is exceeded, the conflicting equipment is moved: .

[0043] Furthermore, the heuristic genetic evolution mechanism in step 4 of the present invention is as follows: Step 4-1: Elite Retention and Heuristic Initialization In standard particle swarm optimization (PSO), the update of an individual's position is guided by the global optimum and the individual's historical optimum, which can easily lead to local optima when dealing with multi-constraint, strongly coupled problems. To enhance global search capabilities and improve population diversity, this paper introduces a heuristic genetic evolution mechanism into the EPSO framework, integrating elite preservation strategies, heuristic initialization, adaptive crossover mutation, and local hill-climbing search to construct a hybrid global jump operator.

[0044] This mechanism is executed once every G=10 generations and applies to all individual particles, achieving co-optimization with the standard PSO process.

[0045] To preserve the current optimal structure, first select the top particles with the highest fitness values ​​in the current particle swarm. Proportion (e.g.) Individuals of this type directly inherit into the new population, forming an elite group:

[0046] in, , The total number of particles, This is the fitness function.

[0047] The elite strategy prevents high-performing individuals from being lost during mutation, ensuring optimization progress. Non-elite individuals are initialized using a heuristic scheduling strategy. They are scheduled sequentially according to their device numbers, with the start time for each time period selected based on the minimum load and fewest conflicts.

[0048] This strategy greedily selects appropriate time periods based on historical scheduling information, reducing the probability of initial unreasonable scheduling.

[0049] Step 4-2, Masking Interleaving Strategy To enhance population diversity and gene recombination efficiency, this paper employs a random masked crossover strategy: 1. Randomly select two parent individuals , ; 2. Construct a random mask Generate offspring:

[0050] That is when When =1, ,when When =0, ,in This indicates bitwise operations.

[0051] This method can preserve the best segments from the parent generation while providing opportunities for structural recombination, effectively preventing premature maturation of the population.

[0052] Step 4-3, Adaptive Probabilistic Mutation Mechanism To maintain the population's exploratory ability, a dynamic mutation probability that varies with the number of iterations is introduced:

[0053] in, This represents the initial mutation probability, with a value such as 0.1. The growth coefficient controls the intensity of variation. This represents the maximum number of iterations. Let be the current iteration algebra.

[0054] The specific mutation operation involves performing mutations on each offspring. If the mutation conditions are met, a gene locus is randomly selected. Add perturbation to its start time

[0055] Mutations can break out of the current local convergence region, increasing population perturbation.

[0056] Step 4-4: Multi-neighbor local hill climbing search For the offspring individuals after crossover mutation, this paper designs a device-by-device neighborhood search mechanism, the specific steps of which are as follows: 1) For each device The current schedule is ; 2) Constructing a local neighborhood: ; 3) Keeping the rest of the equipment unchanged, test the fitness of the new scheduling combination; 4) If the new combination Then update to .

[0057]

[0058] This method is a local search enhancement operation (local hill climbing algorithm), which can effectively improve the quality of offspring individuals and reduce the repair burden.

[0059] In summary, this invention proposes a multi-constraint particle swarm optimization (PSO) method for power maintenance scheduling. It enhances global search capabilities by combining a dynamic inertia weight adjustment mechanism and a midpoint-guided mutation strategy, and introduces a genetic evolution mechanism involving elite retention, random masked crossover, and adaptive probabilistic mutation to improve population diversity and optimization stability. Furthermore, it designs a multi-neighborhood local hill-climbing search and a multi-constraint feasibility repair function to ensure high-quality scheduling schemes while satisfying constraints. Experimental results show that this method outperforms traditional algorithms in terms of convergence speed, solution feasibility, and scheduling quality, demonstrating good engineering applicability and optimization effectiveness.

[0060] Specifically, the beneficial effects of this invention are as follows: 1. Enhanced optimization capabilities: Through dynamic inertia factor and midpoint guidance mechanism, the algorithm's ability to explore in the early stages and converge in the later stages is improved.

[0061] 2. Improved feasibility of solutions: The adaptive repair function ensures that individual particles satisfy all scheduling constraints and effectively generates executable scheduling schemes.

[0062] 3. Enhanced diversity: The introduction of masked crossover and dynamic mutation strategies enhances population diversity and avoids premature maturation. Attached Figure Description

[0063] Figure 1 This is a flowchart illustrating a multi-constraint particle swarm optimization algorithm for power distribution network maintenance scheduling that integrates genetic inspiration and dynamic penalty mechanism according to an embodiment of the present invention. Figure 2 This is a comparison chart of the fitness function curves of the traditional PSO algorithm and the optimized PSO algorithm according to an embodiment of the present invention. Detailed Implementation To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] like Figure 1 As shown in the figure, this embodiment of the invention provides a multi-constraint particle swarm optimization power maintenance scheduling method, which includes the following steps: Step 1: Construct a power distribution network maintenance scheduling optimization model. The optimization model aims to minimize power outage losses and maintenance costs. It considers five types of constraints in the actual maintenance process, namely simultaneous constraints, mutual exclusion constraints, sequential constraints, resource constraints, and time window constraints, forming a multi-objective nonlinear integer programming problem.

[0065] Step 2: Initialize and improve particle swarm optimization parameters, setting parameters such as the number of particles, maximum number of iterations, velocity limit, learning factor, and initial inertia weight. Encode particle positions as maintenance start time vectors, construct the particle swarm population, and initialize the fitness function.

[0066] Step 3: Optimize the distribution network maintenance model using an improved particle swarm optimization algorithm to avoid problems such as getting trapped in local optima and insufficient search diversity that may occur with traditional optimization algorithms, and obtain the optimized distribution network maintenance plan arrangement results.

[0067] Step 3-1: Dynamic inertia weight adjustment mechanism.

[0068] Step 3-2: Midpoint-guided mutation speed update strategy.

[0069] Step 3-3 Feasibility repair function repair().

[0070] Step 4: Introduce a heuristic genetic evolution mechanism into the improved particle swarm framework, and integrate elite retention strategy, heuristic initialization, adaptive crossover mutation and local hill climbing search to construct a hybrid global jump operator.

[0071] Step 4-1: Elite Retention and Heuristic Initialization Step 4-2, Masking Interleaving Strategy Step 4-3, Adaptive Probabilistic Mutation Mechanism Step 4-4: Multi-neighbor local hill climbing search Step 5: After reaching the maximum number of iterations or the convergence condition, output the optimal scheduling particle, decode it into the corresponding equipment maintenance plan, and form an executable scheduling scheme that meets the constraints and optimization objectives.

[0072] Specifically, the model structure in step 1 of the present invention is as follows: The objective function to be optimized is denoted as:

[0073] in, Power outage losses due to maintenance; Maintenance costs for each piece of equipment during the planned period.

[0074] Specifically, it is expressed as follows:

[0075] in, Total number of equipment to be inspected; : Maintenance start time of device i; Maintenance duration of device i (in time period); The economic loss caused by the load interruption due to the maintenance of equipment i in time period t; : The direct economic cost (such as labor, materials, tools, etc.) of equipment i performing maintenance in time period t; Simultaneous constraint: In the maintenance and scheduling of power systems, if two or more devices have control interlocking relationships, are geographically close, or belong to the same operating unit (such as the same line or the same bay), these devices should be maintained within the same time period to avoid repeated power outages, improve maintenance efficiency, and reduce the scope of power outages.

[0076] Therefore, suppose there exists a set of equipment pairs that require simultaneous maintenance:

[0077] Each pair Indicates equipment and Simultaneous maintenance is required. Therefore, the mathematical form of this constraint can be expressed as:

[0078] Mutual Exclusion Constraint: During power distribution network maintenance, some devices may share critical facilities such as protection devices and circuit breakers, as they support the same power supply path within the system structure. If these devices are maintained simultaneously, it may lead to a complete interruption of the power supply path or failure of the protection configuration. Therefore, devices with such logically mutually exclusive relationships must not be maintained simultaneously during overlapping time periods; that is, their maintenance time periods must not overlap at all.

[0079] Let the set of device pairs that need to satisfy mutual exclusion be:

[0080] in Indicates equipment With equipment There is a mutual exclusion relationship, and their maintenance start times are defined as follows: , The maintenance time is as follows: , Then the mathematical expression for the mutual exclusion constraint is:

[0081] Maintenance time window constraints: Each maintenance task must be completed within the specified time window to minimize the impact on the power system.

[0082]

[0083] in and These are the earliest and latest start times for the k-th maintenance item, respectively. This refers to the actual maintenance time.

[0084] Maintenance resource constraints: The manpower, material resources, equipment and financial resources required for maintenance are limited and need to be rationally allocated under the given resource conditions.

[0085]

[0086] in It is the r-th type of resource used in the maintenance plan. It is the total amount of this type of resource.

[0087] Sequence Constraints: In power distribution systems, the maintenance of some critical equipment has sequential dependencies. For example, busbar maintenance can only begin after the circuit breaker has completed its isolation operation, or certain interval maintenance can only be carried out after the main transformer has been completed. In such scenarios, to ensure maintenance safety and standardized work procedures, it is necessary to set sequence constraints for equipment maintenance. Assume there exists a set of equipment pairs that must satisfy the sequence relationship:

[0088] Each pair Indicates equipment Maintenance must precede equipment repair. Finish. :equipment The start time of maintenance; :equipment Maintenance time; :equipment The start time of maintenance.

[0089] In this embodiment, step 3-1 can be implemented by the following method: To achieve a balance between global and local search capabilities, an exponentially decreasing dynamic inertia factor is introduced, defined as follows:

[0090] in, Let be the inertia factor at the t-th iteration. and The maximum and minimum values ​​of the inertia factor are taken in this paper. =0.9, =0.2; The decay rate factor is set to 0.1 in this paper.

[0091] In this embodiment, step 3-2 can be implemented by the following method: Based on the standard PSO velocity update formula, this paper introduces an individual optimal solution. With the global optimal solution The midpoint serves as the guiding center:

[0092] The formulas for updating the particle's velocity and position are:

[0093]

[0094] in, , is a uniformly distributed random number; and The learning factor (usually 2); This is the feasibility repair function. Since the position updates are continuous values, they need to be discretized in the scheduling problem to ensure that the constraints are met.

[0095] In this embodiment, step 3-3 can be implemented by the following method: 1. Legality of maintenance period restored:

[0096] Ensure that the scheduling time does not exceed the limit.

[0097] 2. Simultaneous constraint repair: If devices i and j have simultaneous constraints:

[0098] Directly Assigned value To achieve synchronization.

[0099] 3. Mutual Exclusion Constraint Repair: If the two maintenance periods overlap:

[0100] The adjustment is as follows: Or postpone to the next feasible time. 4. Sequence constraint repair: To ensure that i completes before j:

[0101] If not satisfied, adjust as follows:

[0102] 5. Maximum number of concurrent devices: For example, the maximum number of devices that can be repaired in each time period is Then, the usage for each time period is counted, and if the limit is exceeded, the conflicting equipment is moved:

[0103] In this embodiment, step 4-1 can be implemented by the following method: In standard particle swarm optimization (PSO), the update of an individual's position is guided by the global optimum and the individual's historical optimum, which can easily lead to local optima when dealing with multi-constraint, strongly coupled problems. To enhance global search capabilities and improve population diversity, this paper introduces a heuristic genetic evolution mechanism into the EPSO framework, integrating elite preservation strategies, heuristic initialization, adaptive crossover mutation, and local hill-climbing search to construct a hybrid global jump operator.

[0104] This mechanism is executed once every G=10 generations and applies to all individual particles, achieving co-optimization with the standard PSO process.

[0105] To preserve the current optimal structure, first select the top particles with the highest fitness values ​​in the current particle swarm. Proportion (e.g.) Individuals of this type directly inherit into the new population, forming an elite group:

[0106] in, , The total number of particles, This is the fitness function.

[0107] The elite strategy prevents high-performing individuals from being lost during mutation, ensuring optimization progress. Non-elite individuals are initialized using a heuristic scheduling strategy. They are scheduled sequentially according to their device numbers, with the start time for each time period selected based on the minimum load and fewest conflicts.

[0108] This strategy greedily selects appropriate time periods based on historical scheduling information, reducing the probability of initial unreasonable scheduling.

[0109] In this embodiment, step 4-2 can be implemented by the following method: To enhance population diversity and gene recombination efficiency, this paper employs a random masked crossover strategy: 1. Randomly select two parent individuals , ; 2. Construct a random mask Generate offspring:

[0110] That is when When =1, ,when When =0, ,in This indicates bitwise operations.

[0111] This method can preserve the best segments from the parent generation while providing opportunities for structural recombination, effectively preventing premature maturation of the population.

[0112] In this embodiment, step 4-3 can be implemented by the following method: To maintain the population's exploratory ability, a dynamic mutation probability that varies with the number of iterations is introduced:

[0113] in, This represents the initial mutation probability, with a value such as 0.1. The growth coefficient controls the intensity of variation. This represents the maximum number of iterations. Let be the current iteration algebra.

[0114] The specific mutation operation involves performing mutations on each offspring. If the mutation conditions are met, a gene locus is randomly selected. Add perturbation to its start time

[0115] Mutations can break out of the current local convergence region, increasing population perturbation.

[0116] In this embodiment, step 4-4 can be implemented by the following method: For the offspring individuals after crossover mutation, this paper designs a device-by-device neighborhood search mechanism, the specific steps of which are as follows: 1) For each device The current schedule is ; 2) Constructing a local neighborhood: ; 3) Keeping the rest of the equipment unchanged, test the fitness of the new scheduling combination; 4) If the new combination Then update to .

[0117]

[0118] This method is a local search enhancement operation (local hill climbing algorithm), which can effectively improve the quality of offspring individuals and reduce the repair burden.

[0119] Combination Figure 2As shown, this invention proposes a multi-constraint particle swarm optimization (PSO) method for power maintenance scheduling. It enhances global search capabilities by combining a dynamic inertia weight adjustment mechanism and a midpoint-guided mutation strategy, and introduces a genetic evolution mechanism involving elite retention, random masked crossover, and adaptive probability mutation to improve population diversity and optimization stability. Furthermore, it designs a multi-neighborhood local hill-climbing search and a multi-constraint feasibility repair function to ensure high-quality scheduling schemes while satisfying constraints. Experimental results show that this method outperforms traditional algorithms in terms of convergence speed, solution feasibility, and scheduling quality, demonstrating good engineering applicability and optimization performance.

[0120] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0121] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0122] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the mobile source emission prediction methods based on time-series feature migration described in the above embodiments.

[0123] It is understood that the systems, devices, and storage media provided in the embodiments of the present invention correspond to the methods provided in the embodiments of the present invention, and the explanations, examples, and beneficial effects of the relevant content can be referred to the corresponding parts of the above methods.

[0124] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0125] It should be noted that in this invention, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0126] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0127] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. 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. Such 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 multi-constraint particle swarm optimization method for power maintenance scheduling, characterized in that, Includes the following steps: Step 1: Construct a power distribution network maintenance scheduling optimization model. The optimization model aims to minimize power outage losses and maintenance costs. It considers five types of constraints in the actual maintenance process: simultaneous constraints, mutual exclusion constraints, sequential constraints, resource constraints, and time window constraints, forming a multi-objective nonlinear integer programming problem. Step 2: Initialize the improved particle swarm optimization parameters, setting parameters such as the number of particles, maximum number of iterations, velocity limit, learning factor, and initial inertia weight; encode the particle positions as maintenance start time vectors, construct the particle swarm population, and initialize the fitness function; Step 3: The improved particle swarm optimization algorithm is used to optimize the distribution network maintenance scheduling optimization model to obtain the optimized distribution network maintenance plan arrangement results; Step 4: Introduce a heuristic genetic evolution mechanism into the improved particle swarm framework, and integrate elite preservation strategy, heuristic initialization, adaptive crossover mutation and local hill climbing search to construct a hybrid global jump operator; Step 5: After reaching the maximum number of iterations or the convergence condition, output the optimal scheduling particle, decode it into the corresponding equipment maintenance plan, and form an executable scheduling scheme that meets the constraints and optimization objectives.

2. The multi-constraint particle swarm optimization power maintenance scheduling method according to claim 1, characterized in that, The structure of the power distribution network maintenance and scheduling optimization model in step S1 is as follows: The objective function to be optimized is denoted as: in, Power outage losses due to maintenance; Maintenance costs for each piece of equipment during the planned period; Specifically, it is expressed as follows: in, Total number of equipment to be inspected; : Maintenance start time of device i; The maintenance duration of equipment i, in time intervals; The economic loss caused by the load interruption due to the maintenance of equipment i in time period t; The direct economic cost of equipment i performing maintenance in time period t; Simultaneous constraint: In the maintenance and scheduling of the power system, if two or more devices have control interlocking relationships, are geographically close, or belong to the same operating unit during operation, these devices should be maintained within the same time period. Therefore, suppose there exists a set of equipment pairs that require simultaneous maintenance: Each pair Indicates device and If simultaneous maintenance is required, then the mathematical form of this constraint is: ； Mutual exclusion constraint: During the maintenance of the power distribution network, some equipment may be completely interrupted or the protection configuration may fail if they are maintained at the same time because they support the same power supply path in the system structure or share critical facilities such as protection devices and circuit breakers. Let the set of device pairs that need to satisfy mutual exclusion be: in Indicates device With equipment There is a mutual exclusion relationship, and their maintenance start times are defined as follows: , The maintenance time is as follows: , Then the mathematical expression for the mutual exclusion constraint is: Maintenance time window constraints: Each maintenance task must be completed within the specified time window to minimize the impact on the power system; in and These are the earliest and latest start times for the k-th maintenance item, respectively. This refers to the actual maintenance time. Maintenance resource constraints: The manpower, material resources, equipment, and financial resources required for maintenance are limited and need to be rationally allocated within the constraints of available resources; in It is the r-th type of resource used in the maintenance plan. It refers to the total amount of this type of resource; Sequence Constraints: Set the order constraints for equipment maintenance; assume there exists a set of equipment pairs that need to satisfy the sequence relationship: Each pair Indicates device Maintenance must precede equipment repair. Finish; :equipment The start time of maintenance; :equipment Maintenance time; :equipment The start time of maintenance.

3. The multi-constraint particle swarm optimization power maintenance scheduling method according to claim 1, characterized in that, The improved particle swarm optimization algorithm in step 3 is as follows: Step 3-1: Dynamic Inertia Weight Adjustment Mechanism An exponentially decreasing dynamic inertia factor is introduced, defined as follows: in, Let be the inertia factor at the t-th iteration. and These represent the maximum and minimum values ​​of the inertia factor. This is the decay rate factor; Step 3-2, Midpoint-guided mutation rate update strategy; Based on the standard PSO velocity update formula, an individual optimal solution is introduced. With the global optimal solution The midpoint serves as the guiding center: The formulas for updating the particle's velocity and position are: in, , is a uniformly distributed random number; and For learning factors; This is a feasibility repair function; since the position update is a continuous value, it needs to be discretized in the scheduling problem and ensure that the constraints are met. Step 3-3 Feasibility Repair Function repair() 1) Restoration of the legality of the maintenance period: Ensure that the scheduling time does not exceed the limit; 2) Simultaneous constraint repair: If devices i and j have simultaneous constraints: Directly Assigned value Achieve synchronization; 3) Mutual exclusion constraint repair: If the two maintenance periods overlap: The adjustment is as follows: Or postpone it to the next feasible time; 4) Sequence constraint repair: To ensure that i completes before j: If not satisfied, adjust as follows: 5) Maximum number of concurrent devices: For example, the maximum number of devices that can be repaired in each time period is Then, the usage for each time period is counted, and if the limit is exceeded, the conflicting equipment is moved: 。 4. The multi-constraint particle swarm optimization power maintenance scheduling method according to claim 3, characterized in that, The heuristic genetic evolution mechanism in step 4 is as follows: Step 4-1: Elite Retention and Heuristic Initialization A heuristic genetic evolution mechanism is introduced into the EPSO framework, which integrates elite preservation strategy, heuristic initialization, adaptive crossover mutation and local hill climbing search to construct a hybrid global jump operator; This mechanism is executed once every G=10 generations and applies to all individual particles, achieving co-optimization with the standard PSO process; To preserve the current optimal structure, first select the top particles with the highest fitness values ​​in the current particle swarm. A certain proportion of individuals are directly inherited into the new population, forming an elite group: in, , The total number of particles, The fitness function; Non-elite devices are initialized using a heuristic scheduling strategy; scheduling is performed sequentially according to device number, and the start time with the least load and fewest conflicts is selected for each time period. This strategy greedily selects suitable time periods based on historical scheduling information, reducing the probability of initial unreasonable scheduling; Step 4-2, Masking Interleaving Strategy A random masking crossover strategy is adopted: 1) Randomly select two parent individuals , ; 2) Construct a random mask Generate offspring: That is when When =1, ,when When =0, ,in Indicates bitwise operation; Step 4-3, Adaptive Probabilistic Mutation Mechanism Introducing a dynamic mutation probability that varies with the number of iterations: in, This represents the initial mutation probability, with a value such as 0.

1. The growth coefficient controls the intensity of variation. This represents the maximum number of iterations. Let this be the current iteration algebra; The specific mutation operation involves performing mutations on each offspring. If the mutation conditions are met, a gene locus is randomly selected. Add perturbation to its start time Mutations can break out of the current local convergence region, increasing population perturbation.

5. The multi-constraint particle swarm optimization power maintenance scheduling method according to claim 1, characterized in that, Step 4 also includes, Step 4-4: Multi-neighbor local hill climbing search For the offspring individuals after crossover mutation, a device-by-device neighborhood search mechanism was designed, with the following specific steps: 1) For each device The current schedule is ; 2) Constructing a local neighborhood: ; 3) Keeping the rest of the equipment unchanged, test the fitness of the new scheduling combination; 4) If the new combination Then update to . 。 6. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, the processor performs the steps of the method as described in any one of claims 1 to 6.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the computer program is executed by the processor, the processor performs the steps of the method as described in any one of claims 1 to 6.

Images