A power system operator scheduling optimization method, system, device and medium

Abstract

The present application relates to the technical field of power operation scheduling, and provides a power system operation personnel scheduling optimization method, system, device and medium, which comprises obtaining power operation task data, scheduling constraint rules, historical scheduling data and personnel skill data; feature extraction is performed on the historical scheduling data to obtain a case feature data set, and an expert preference discrimination model is constructed according to the case feature data set; according to the power operation task data, task period decomposition is performed to obtain a target shift combination with the minimum total task operation cost as the optimization objective; and according to the target shift combination, the scheduling constraint rules, the personnel skill data and the expert preference discrimination model, personnel allocation optimization is performed based on a cultural gene algorithm to obtain a target scheduling scheme. The present application can ensure the scheduling optimization quality and efficiency, take into account the scheduling scheme feasibility and personnel allocation efficiency, and improve the practicality and reliability of the scheduling scheme based on the two-stage optimization mechanism of task period decomposition first and then personnel allocation.

Description

Technical Field

[0001] This invention relates to the field of power operation scheduling technology, and in particular to a method, system, equipment and medium for optimizing the scheduling of power system operators. Background Technology

[0002] Power system operations are highly complex, often involving multiple substations and various types of maintenance tasks, requiring maintenance personnel to possess different professional skills, which poses a significant challenge to the reasonable scheduling of operations personnel.

[0003] Existing task scheduling technologies mainly include rule-based, mathematical programming-based, and traditional genetic algorithm-based scheduling schemes. However, their application in power system operation scheduling has significant limitations: rule-based scheduling schemes require the use of pre-set business rules, and their implementation heavily relies on the writing and maintenance of rule bases based on domain expert experience. This makes it difficult to consider multiple constraints such as personnel skill matching, geographical distribution, and work duration balance. Furthermore, they cannot adaptively learn effective patterns from historical scheduling and dispatcher preferences, resulting in low scheduling efficiency, low personnel utilization, and inflexible scheduling. Mathematical programming-based scheduling schemes typically model the scheduling problem as a mixed-integer linear programming problem. The scheduling results obtained by optimization rely heavily on the linear assumptions of the problem and are difficult to accurately express nonlinear factors such as personnel preferences and team collaboration efficiency. Only when the model is accurate and the data is complete can the mathematically optimal solution be obtained. Moreover, it cannot meet the scheduling timeliness requirements when the problem scales up. The scheduling scheme based on the traditional genetic algorithm encodes the scheduling scheme into chromosomes and performs genetic operations to perform iterative optimization. When dealing with complex constraint scenarios in maintenance operation scheduling, it is prone to generating a large number of infeasible solutions. It also relies on complex repair mechanisms or penalty functions to handle constraints, which not only has a large computational cost, but may also cause the search process to get stuck in the suboptimal solution region, making it difficult to guarantee the actual feasibility of the scheduling scheme. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method, system, equipment, and medium for optimizing the scheduling of power system maintenance personnel, so as to efficiently and reliably generate practical maintenance personnel scheduling schemes that meet the hard constraints of substation maintenance scenarios, conform to expert experience, and are interpretable.

[0005] In a first aspect, embodiments of the present invention provide a method for optimizing the scheduling of power system workers, the method comprising: Acquire power operation task data, scheduling constraint rules, historical scheduling data, and personnel skill data; the historical scheduling data includes a success case database, an expert experience database, and a problem case database. Feature extraction is performed on the historical scheduling data to obtain the corresponding case feature dataset, and an expert preference discrimination model is constructed based on the case feature dataset; Based on the power operation task data, the task time period is decomposed with the goal of minimizing the total cost of the task operation, and the target shift combination is obtained. Based on the target shift combination, the shift scheduling constraints, the personnel skill data, and the expert preference discrimination model, personnel allocation is optimized using the cultural gene algorithm to obtain the target shift scheduling scheme.

[0006] Furthermore, the case feature dataset includes case features and corresponding scheduling effect labels for each scheduling case; the case features include personnel features and scheduling features; the personnel features include the skill matching degree, task work experience, regional familiarity, time matching degree, workload, and performance score of each operator; the scheduling features include scheduling experience matching degree and personnel utilization efficiency; the scheduling effect labels include high-quality scheduling labels and low-quality scheduling labels.

[0007] Furthermore, the step of constructing an expert preference discrimination model based on the case feature dataset includes: The expert preference discrimination model is constructed by training the decision tree algorithm using the case features corresponding to the high-quality scheduling label and the low-quality scheduling label in the case feature dataset as positive and negative samples, respectively.

[0008] Furthermore, the step of decomposing the task time periods based on the power operation task data, with the optimization objective of minimizing the total cost of the task operation, to obtain the target shift combination includes: With minimizing the total cost of the task as the optimization objective, a task decomposition optimization function is constructed; Based on the task decomposition optimization function, a task decomposition optimization model is constructed based on preset optimization constraints; the preset optimization constraints include daily working hours constraints and total equipment maintenance requirements constraints. Based on the daily working hour constraints and the power operation task data, shifts are enumerated to obtain a candidate shift set; Based on the candidate shift set, the task decomposition optimization model is equivalently transformed to obtain the shift combination optimization model; The target train schedule combination is obtained by solving the optimization model of the train schedule combination.

[0009] Further, the step of optimizing personnel allocation based on the cultural gene algorithm according to the target shift combination, the scheduling constraint rules, the personnel skill data, and the expert preference discrimination model to obtain the target shift plan includes: Based on the target shift combination, the scheduling constraint rules, and the personnel skill data, initialize the scheduling scheme population using heuristic rules; Based on the scheduling scheme population and the dual-drive fitness function, personnel allocation is optimized using the cultural gene algorithm to obtain the target scheduling scheme; the dual-drive fitness function is obtained by weighted fusion of the expert preference probability and normalized time cost of the scheduling scheme.

[0010] Further, the step of optimizing personnel allocation based on the cultural gene algorithm according to the scheduling scheme population and the preset dual-drive fitness function to obtain the target scheduling scheme includes: Based on the effective working time and total task delay of each initial scheme in the scheduling scheme population, a total time cost analysis is performed to obtain the corresponding normalized time cost. The scheme characteristics of each initial scheme are analyzed according to the expert preference discrimination model to obtain the corresponding expert preference probability; Based on the normalized time cost and expert preference probability of each initial scheme, the corresponding comprehensive fitness value is calculated based on the dual-drive fitness function, and multiple parent schemes are obtained based on the comprehensive fitness value. Each parent scheme is subjected to crossover and mutation operations in turn to obtain multiple child schemes, and the optimal child scheme is obtained based on all the child schemes. The scheduling scheme population is updated based on all the parent schemes and all the optimal child schemes. The optimal scheme of the updated scheduling scheme population is obtained, and when the updated scheduling scheme population meets the preset iteration termination condition, the optimal scheme is taken as the target scheduling scheme.

[0011] Furthermore, the method also includes: A target scheduling table is generated based on the target scheduling scheme, and a comprehensive evaluation of the target scheduling table is conducted based on a preset indicator system to obtain the corresponding scheduling evaluation results; the preset indicator system includes task completion rate, resource utilization rate, skill matching degree, cost-effectiveness and similarity of high-quality scheduling.

[0012] Secondly, embodiments of the present invention provide a power system operator scheduling optimization system, the system comprising: The data acquisition module is used to acquire power operation task data, scheduling constraint rules, historical scheduling data, and personnel skill data; the historical scheduling data includes a success case library, an expert experience library, and a problem case library. The model building module is used to extract features from the historical scheduling data to obtain the corresponding case feature dataset, and to build an expert preference discrimination model based on the case feature dataset. The task decomposition module is used to decompose the task time period based on the power operation task data, with the optimization objective of minimizing the total cost of the task operation, to obtain the target shift combination. The personnel allocation module is used to optimize personnel allocation based on the cultural gene algorithm according to the target shift combination, the scheduling constraint rules, the personnel skill data and the expert preference discrimination model, so as to obtain the target shift plan.

[0013] Thirdly, embodiments of the present invention also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0014] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0015] This invention provides a method, system, computer equipment, and storage medium for optimizing the scheduling of power system workers. The method acquires power operation task data, scheduling constraints, historical scheduling data including a success case library, an expert experience library, and a problem case library, and personnel skill data. It then extracts features from the historical scheduling data to obtain corresponding case feature datasets. Based on these datasets, it constructs an expert preference discrimination model. Using the power operation task data, it decomposes the task time periods to obtain target shift combinations with the optimization objective of minimizing the total task operation cost. Finally, based on the target shift combinations, scheduling constraints, personnel skill data, and the expert preference discrimination model, it optimizes personnel allocation using a cultural gene algorithm to obtain the target scheduling scheme. Compared with existing technologies, this power system worker scheduling optimization method, based on a two-stage optimization mechanism of first decomposing the task time periods and then allocating personnel, ensures the quality and efficiency of scheduling optimization while considering the feasibility of the scheduling scheme and the efficiency of personnel allocation, thus improving the practicality and reliability of the scheduling scheme. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the method for optimizing the scheduling of power system workers in an embodiment of the present invention; Figure 2 This is a schematic diagram of the power system operator scheduling optimization system in an embodiment of the present invention; Figure 3 This is an internal structural diagram of the computer device in an embodiment of the present invention; The attached figures are labeled as follows: 1. Data acquisition module; 2. Model building module; 3. Task decomposition module; 4. Personnel allocation module. Detailed Implementation

[0017] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. Obviously, the embodiments described below are only part of the embodiments of this invention and are used to illustrate the invention, but are not intended to limit the scope of the 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.

[0018] The power system operator scheduling optimization method provided by this invention can be understood as addressing the current application situation where existing scheduling methods struggle to simultaneously achieve both scheduling feasibility and optimal personnel allocation efficiency. This method proposes a two-stage scheduling optimization approach: first, it decomposes task time periods using mixed-integer linear programming to obtain target shift combinations that satisfy basic working hours and resource constraints; then, it uses a machine learning-enhanced cultural gene algorithm for personnel allocation. This ensures the practical feasibility of the scheduling plan and the rationality of the dispatching. The following embodiments will provide a detailed description of the power system operator scheduling optimization method of this invention.

[0019] In one embodiment, such as Figure 1 As shown, a method for optimizing the scheduling of power system operators is provided, including: S11. Obtain power operation task data, scheduling constraints, historical scheduling data, and personnel skill data; among which, power operation task data can be understood as the basic information of the actual maintenance tasks to be performed, which may include the number of power operation days, the number of substations to be operated, the type of equipment to be maintained, the geographical location and voltage level of each substation, and the maintenance costs, maintenance time, and skill requirements for different types of equipment within the substation; scheduling constraints may include skill matching requirements and priority rules for different equipment maintenance, the maximum total working hours limit for the work day, personnel work preferences (including each person's preferred work area, willingness to work overtime or travel), and team collaboration habits ( (Does a fixed, efficient personnel pairing exist?) Historical scheduling data can be understood as a database of scheduling schemes obtained by combining historical scheduling schemes and expert scheduling experience. This may include a success case database (a database of high-quality scheduling schemes with excellent execution results), an expert experience database (a database of scheduling schemes specified based on expert experience), and a problem case database (a database of poor-performing scheduling schemes). Personnel skill data may include basic information of each assignable personnel (including name, work group, and position), skill qualifications (including skill list, qualification certificates, and skill levels), and work habits (including current work status, current available time, current workload, and historical task completion quality). It should be noted that power operation task data, scheduling constraint rules, historical scheduling data, and personnel skill data can all be obtained from relevant information databases of the power system based on the optimized scheduling of personnel for the tasks to be performed; these will not be detailed here.

[0020] S12. Extract features from the historical scheduling data to obtain the corresponding case feature dataset, and construct an expert preference discrimination model based on the case feature dataset. The case feature dataset can be understood as a dataset composed of features affecting the execution effect of all scheduling cases obtained from analyzing all scheduling cases in the historical scheduling data. That is, the case feature dataset includes the case features of each scheduling case and the corresponding scheduling effect label. To ensure the comprehensiveness and reliability of expert preference prediction analysis based on case features, this embodiment preferably sets the case features to include personnel features and scheduling features. Personnel features include the skill matching degree, task work experience, regional familiarity, time matching degree, workload, and performance score of each operator. Scheduling features include scheduling experience matching degree and personnel utilization efficiency. The corresponding scheduling effect labels include high-quality scheduling labels and low-quality scheduling labels.

[0021] In this embodiment, scheduling schemes labeled as "high-quality scheduling" can be understood as those that have appeared repeatedly in historical scheduling data and have already been implemented. Conversely, scheduling schemes labeled as "low-quality scheduling" can be understood as those that were initially arranged but ultimately cancelled in historical scheduling data. To facilitate the expert preference discrimination model in comprehensively and effectively capturing the success patterns involved in high-quality scheduling schemes, this embodiment preferably uses all cases in the success case library and expert experience library of historical scheduling data as high-quality scheduling cases, and the problem case library as low-quality scheduling cases. Then, case features are extracted from each scheduling case, and each case feature is labeled with either a high-quality scheduling label or a low-quality scheduling label to obtain the required case feature dataset. It should be noted that each scheduling case includes task data, scheduling constraints, personnel skill data, and the corresponding scheduling scheme. Specifically, the case feature extraction process for each scheduling case is as follows: 1) Skill matching degree can be understood as the degree of matching between the skill level of the operator and the skill level required by the task. It is used to measure whether the assigned operator can complete the relevant task. In practical application, the skill level can be quantified into the corresponding level score first. Then, the skill matching degree of different operators in the current scheduling plan can be obtained by comparing the skill level of each operator in the scheduling plan with the corresponding level score of the task required skill level. 2) Task work experience can be understood as the amount of experience accumulated by operators in completing similar tasks, which is used to measure whether operators have the ability to handle related tasks. In practical applications, based on the task type assigned to operators in the current scheduling plan, the ratio of the number of times operators have completed this type of task in the current scheduling plan and all previous scheduling plans to the total number of this type of task in the current scheduling plan and all previous scheduling plans can be calculated to obtain the task work experience of operators in the current scheduling plan. 3) Geographical familiarity can be understood as the degree to which operators are familiar with the substation area they are working on. It is used to measure whether operators can reduce travel time and the risk of misoperation. In practical applications, based on the substation area where the operator is assigned a task in the current shift plan, the ratio of the number of times the operator completes tasks in that substation area in the current shift plan and all previous shift plans to the total number of tasks completed by the operator in the current shift plan and all previous shift plans can be calculated to obtain the operator's geographic familiarity in the current shift plan. 4) Time matching degree can be understood as the degree of matching between the available time of the operator and the task time window, which is used to measure the likelihood of time conflict in task scheduling. In practical applications, the overlap length between the available time of the operator before executing the current scheduling plan and the task time in the current scheduling plan can be obtained, and then the ratio of the overlap length to the task time length can be calculated to obtain the time matching degree of the operator in the current scheduling plan. 5) Workload status can be understood as the degree of workload that operators bear when executing the current shift schedule, and is used to measure the balance of personnel workload; in practical applications, the corresponding workload status can be obtained by calculating the ratio of the task load of operators when executing the current shift schedule to the maximum capacity of operators. 6) Performance rating can be understood as a comprehensive evaluation of the historical task completion quality of operators, used to measure the reliability of operators' task execution; in practical applications, the comprehensive ranking of the task completion quality of all shift plans before the current shift plan can be obtained, and then the corresponding performance rating can be obtained based on the ratio of the comprehensive ranking to the total number of operators. 7) Scheduling experience matching degree can be understood as the overall assessment of the experience matching degree of the scheduling plan, used to measure the inheritance of high-quality experience in the scheduling plan. In practical application, the group allocation mode of each task in the current scheduling plan can be compared with the corresponding scheduling mode of the historical high-quality scheduling plan. If the group allocation mode of the current task appears in the historical high-quality scheduling plan, the number of matching tasks between the scheduling plan and the historical high-quality scheduling plan is increased by 1 until the group allocation mode of all tasks is compared to obtain the total number of matching tasks between the current scheduling plan and the historical high-quality scheduling plan. Then, the total number of matching tasks is compared with the total number of tasks corresponding to the current scheduling plan to obtain the required scheduling experience matching degree. 8) Personnel utilization efficiency can be understood as the degree of utilization of operational human resources in the scheduling plan from an overall perspective. It is used to measure the level of human resource costs in the scheduling plan (overall utilization efficiency of operational personnel). In practical applications, the total effective working time and total available time of the current scheduling plan can be obtained by summing the estimated working hours of the tasks assigned to each shift in the current scheduling plan and the maximum working hours that each shift can theoretically be assigned. Then, by comparing the total effective working time with the total available time, the required personnel utilization efficiency can be obtained.

[0022] By extracting features from each scheduling case in historical scheduling data to obtain personnel features and scheduling features, a case feature dataset construction mechanism can be established. This mechanism enables multi-dimensional feature extraction and analysis of each scheduling case from the micro-dimensional dimension of individual worker capabilities and the macro-dimensional dimension of scheduling scheme combination. This not only ensures the reliability, rationality, and comprehensiveness of case feature extraction but also ensures the diversity of case features, providing a reliable and comprehensive training data foundation for the high-quality training and construction of subsequent expert preference discrimination models.

[0023] The expert preference discrimination model in this embodiment can be understood as a machine learning model that analyzes the case features of scheduling schemes and outputs the probability value of a scheduling scheme belonging to a high-quality scheduling scheme. In principle, the expert preference discrimination model can be based on any model capable of binary classification as the foundation model for training. However, considering the complexity and diversity of case feature data in the case feature dataset and the nonlinear relationships and interactions between case features, in order to reduce the workload of feature engineering and obtain a reliable expert preference discrimination model simply and efficiently, this embodiment preferably uses a decision tree algorithm to construct the expert preference discrimination model. Specifically, the steps of constructing the expert preference discrimination model based on the case feature dataset include: The expert preference discrimination model is constructed by training the decision tree algorithm using the case features corresponding to the high-quality scheduling label and the low-quality scheduling label in the case feature dataset as positive and negative samples, respectively. It should be noted that the specific process of training and constructing the expert preference discrimination model can be implemented by referring to the existing decision tree algorithm training and construction methods, which will not be described in detail here.

[0024] This embodiment builds a model mechanism based on the feature patterns of scheduling schemes extracted from historical scheduling data. This mechanism can score scheduling scheme preferences like an expert. It not only effectively ensures the coverage of training samples and reduces the risk of overfitting by leveraging the diversity of case feature samples in the case feature dataset, but also enables the model to accurately learn the feature differences between high-quality and low-quality scheduling schemes based on the binary classification labels of case features in the case feature dataset. This transforms the decision rules implicit in the task allocation of high-quality scheduling schemes into interpretable and reusable classification rules, forming a reliable discrimination boundary for expert preferences. This allows for the systematic extraction of high-quality scheduling experience, ensuring the efficiency and reliability of the expert preference discrimination model construction, and providing reliable guidance for subsequent personnel allocation optimization.

[0025] S13. Based on the power operation task data, the task time period is decomposed with the goal of minimizing the total cost of the task operation to obtain a target shift combination. The target shift combination can be understood as a globally feasible task execution schedule obtained by decomposing the power operation task into time periods based on working hours and resource constraints, according to the power operation task data. To ensure the reliability of obtaining the target shift combination, this embodiment preferably uses the goal of minimizing the total cost of the task operation and performs task time period decomposition through mixed-integer linear programming. Specifically, the step of decomposing the task time period based on the power operation task data with the goal of minimizing the total cost of the task operation to obtain the target shift combination includes: With minimizing the total cost of the task as the optimization objective, a task decomposition optimization function is constructed. The total cost of the task is derived from a comprehensive analysis of the maintenance costs for different types of equipment at different substations on each work day, based on the number of power operation days, the number of substations to be operated, and the types of equipment required for maintenance. The corresponding task decomposition optimization function can be expressed as follows: in, For the number of days of power operation; The number of substations awaiting operation; For the number of equipment types; For the first Substation operating day Complete equipment maintenance The number of task items; For the first Substation operating day Maintenance equipment Maintenance costs.

[0026] Based on the task decomposition optimization function, a task decomposition optimization model is constructed based on preset optimization constraints; wherein, the preset optimization constraints include daily working hours constraints and total equipment maintenance requirements constraints. 1) The daily working hour constraint can be understood as follows: the total working hours required for all maintenance tasks of all equipment in all substations on each workday cannot exceed the maximum available total working hours limit for the corresponding workday, which can be expressed as: in, For equipment maintenance Maintenance time; For the first The maximum available total working hours per workday; 2) The constraint on total equipment maintenance demand can be understood as follows: for each type of maintenance equipment within a substation, the total number of equipment scheduled for maintenance during the entire power operation planning period must be greater than or equal to the total maintenance demand for that type of equipment at that substation. This can be expressed as: in, For the first Substation operating day Maintenance equipment The number of maintenance needs.

[0027] Based on the daily working hour constraints and the power operation task data, shifts are enumerated to obtain a candidate shift set. The candidate shift set can be understood as a set of feasible shifts that meet the basic working hour requirements by enumerating all possible task time period decomposition methods based on the daily working hour constraints and power operation task data, which can provide a basic candidate space for subsequent refined task personnel allocation. In practical applications, the estimated maintenance time of each task item can be extracted from the power operation task data, and the maximum available total working hours limit for the work day can be obtained. Then, based on the tasks to be executed, a state search tree is constructed, a search stack is initialized to store the sequence of currently explored task nodes, and a cumulative duration variable is defined. Based on a combinatorial generation algorithm (such as depth-first search), the task nodes are traversed. Each time a new task node is pushed onto the search stack, its corresponding estimated maintenance time is added to the cumulative duration variable in real time. In this process, a pruning mechanism can be introduced. The cumulative duration variable is compared with the corresponding maximum available total working hours limit in real time through conditional judgment instructions. If the total duration of each work day is less than or equal to the corresponding maximum available total working hours limit, the task sequence in the current stack is converted into an array format and stored as a candidate set as a feasible shift. If it exceeds the maximum available total working hours limit, pruning is performed, triggering a backtracking operation, popping the top task node from the stack and deducting its duration simultaneously, and stopping the subsequent combination of that branch. Through the above-mentioned traversal mechanism with constraint pruning, the required candidate shift set can be obtained efficiently.

[0028] Based on the candidate shift set, the task decomposition optimization model is equivalently transformed to obtain a shift combination optimization model; wherein, the shift combination optimization model is understood as an optimization model obtained by converting the task decomposition optimization model established from the perspective of task allocation into the perspective of shift selection, and the specific implementation process includes: The task decomposition optimization function in the task decomposition optimization model, which obtains the total task cost by summing the costs of each task, is transformed into an optimization objective function that first summarizes the task costs by shift to obtain the shift cost, and then summarizes the costs of each shift to obtain the total task cost. That is, the sum of the costs of all selected shifts in the shift combination optimization model is exactly the same as the sum of the costs of all tasks in the task decomposition optimization function, which can be expressed as: in, For candidate class schedules; For the candidate classes, the first b The cost of a shift is the total cost of the tasks included in that shift. For the first b The selection variables corresponding to each class, when When the value is 1, it indicates that the scheduled train has been selected. A value of 0 indicates that the scheduled shift is not selected.

[0029] When performing an equivalent transformation on the preset optimization constraints in the task decomposition optimization model, since the daily working hour constraint has already been reflected in the generation of the candidate shift set, only the total equipment maintenance demand constraint needs to be transformed, which can be expressed as: In the formula, For the candidate classes, the first b Substations on each shift Complete equipment maintenance The number of task items; For substation Maintenance equipment The number of maintenance needs.

[0030] It should be noted that by using the above methods to convert the task decomposition optimization model into an equivalent shift combination optimization model, the optimization variable is essentially changed from task allocation to shift selection while keeping the physical meaning unchanged. The coarse-grained shift variable replaces the fine-grained task variable, thereby significantly reducing the complexity of the optimization model and facilitating efficient solution.

[0031] The shift combination optimization model is solved to obtain the target shift combination; wherein, the shift combination optimization model is a mixed integer linear programming model, which can directly call the existing optimization solver for precise optimization, and finally obtain the globally optimal shift combination that simultaneously satisfies the basic working time constraints and resource constraints and has the lowest total cost as the target shift combination.

[0032] This embodiment is based on the principle of integer programming. It takes minimizing the total cost of task operations as the optimization objective and introduces an optimization modeling mechanism with daily working hours and total equipment maintenance requirements as constraints. It can perform cost-optimized task time decomposition optimization of power operation tasks based on hard constraints, so as to efficiently and accurately obtain the global optimal shift combination and provide a reliable analytical basis for subsequent personnel allocation optimization.

[0033] S14. Based on the target shift combination, the scheduling constraint rules, the personnel skill data, and the expert preference discrimination model, personnel allocation optimization is performed using a cultural gene algorithm to obtain a target scheduling scheme. The target scheduling scheme can be understood as the optimal task scheduling strategy obtained by using a machine learning-enhanced cultural gene algorithm to optimize personnel allocation for all shifts in the target shift combination. To ensure the efficiency of obtaining the target scheduling scheme while also inheriting the pattern characteristics of existing high-quality scheduling schemes, and to ensure the practical feasibility and scheduling rationality of the target scheduling scheme, this embodiment preferably uses an expert preference discrimination model based on machine learning that can mine excellent scheduling features in the scheduling scheme as a soft constraint guide in personnel allocation optimization. This model is combined with time cost and worker skill matching to enhance and optimize the cultural gene algorithm, thereby achieving personnel allocation optimization for the target shift combination.

[0034] Specifically, the step of optimizing personnel allocation based on the cultural gene algorithm according to the target shift combination, the shift scheduling constraint rules, the personnel skill data, and the expert preference discrimination model to obtain the target shift scheduling scheme includes: Based on the target shift combination, the scheduling constraints, and the personnel skill data, a population of scheduling schemes is initialized using heuristic rules. This population can be understood as a set of multiple complete task-personnel allocation methods generated for the target shift combination based on heuristic rules. The acquisition process may include: first, obtaining an initial personnel allocation scheme corresponding to the target shift combination as the first individual (initial scheme) based on the scheduling constraints and personnel skill data, ensuring that a basic feasible solution is retained in the population; second, under the conditions of the scheduling constraints, randomly exchanging personnel allocations between different tasks within the same time period, or randomly selecting other personnel with corresponding skills and available time for a certain task, generating multiple new personnel allocation schemes to increase population diversity and obtain the required population of scheduling schemes, providing a rich initial solution space for subsequent optimization search. It should be noted that the task allocation personnel for each shift in the initialized population of scheduling schemes must meet the requirement that the task allocation personnel possess the skills required to perform the shift task, ensuring that the initial solution naturally satisfies the hard requirements of skills and time period, thus guaranteeing the physical feasibility of the initial solution.

[0035] Based on the scheduling scheme population and the dual-drive fitness function, personnel allocation optimization is performed using the cultural gene algorithm to obtain the target scheduling scheme. The dual-drive fitness function can be understood as a comprehensive analysis function used to optimize and evaluate individuals in the scheduling scheme population during the iterative optimization process of the cultural gene algorithm. To achieve refined personnel matching and scheduling optimization, while ensuring the efficiency of obtaining the target scheduling scheme and its application feasibility, it also enables efficient allocation of human resources. Preferably, in this embodiment, the dual-drive fitness function is constructed based on a weighted fusion of the expert preference probability and normalized time cost of the scheduling scheme, and can be expressed as: In the formula, in, For the current scheduling plan Middle shift Effective working hours of indoor workers; For the current scheduling plan Middle shift Total task delay time; and The current shift scheduling plan Middle shift The deadline and estimated completion time for the task; and The weighting coefficients for effective working time and total task delay time can be set based on actual application needs to meet [the requirements]. ; This represents the original time cost value of the current scheduling plan; and These are the maximum and minimum original time cost values ​​for all individuals in the current generation of the scheduling scheme population, respectively. and These represent the normalized time cost of the scheduling scheme and the probability of expert preference, respectively, and the probability of expert preference... Based on the expert preference discrimination model, it directly reflects the degree of fit of the scheduling scheme to the high-quality scheduling experience in the nonlinear dimension; and These are the weights for expert preference probability and normalized time cost, respectively, which can be set based on application scenario requirements to meet [the needs of the application]. ; The fitness value is calculated based on the dual-drive fitness function.

[0036] The expert preference probability in the dual-drive fitness function provided in this embodiment reflects the nonlinear fitting of historical high-quality scheduling experience, which can guide the optimization algorithm to evolve in a direction that conforms to actual work habits. Furthermore, the normalized time cost can quantify the efficiency performance of the scheduling scheme in terms of time utilization and timely task completion. The weighted fusion of the two avoids local optima that may be caused by relying solely on historical experience, and also prevents the pursuit of time efficiency at the expense of the feasibility and safety of actual operations. In other words, compared with the existing single evaluation method that only relies on the objective function, it can ensure that high-fitness scheduling schemes not only perform well in objective indicators, but also implicitly contain patterns that conform to high-quality scheduling experience, thereby avoiding the loss of population diversity that may be caused by hard elimination. At the same time, it also solves the problem that linear systems cannot represent complex high-quality scheduling experience.

[0037] After determining the aforementioned dual-drive fitness function, genetic operations can be performed based on the scheduling scheme population and the cultural gene algorithm to obtain the target scheduling scheme. To maintain individual diversity within the scheduling scheme population during iterative optimization, this embodiment preferably employs genetic operations including tournament mode selection, single-point crossover, and uniform mutation to optimize personnel allocation for the target shift combination. Specifically, the step of optimizing personnel allocation based on the cultural gene algorithm according to the scheduling scheme population and the preset dual-drive fitness function to obtain the target scheduling scheme includes: Based on the effective working hours and total task delay time of each initial scheme in the aforementioned scheduling scheme population, a total time cost analysis is performed to obtain the corresponding normalized time cost; wherein, the calculation process of the normalized actual cost can be based on the normalized time cost in the aforementioned dual-drive fitness function. The calculation formula is obtained, and will not be repeated here.

[0038] The scheme characteristics of each initial scheme are analyzed according to the expert preference discrimination model to obtain the corresponding expert preference probability. That is, after generating each initial scheme, the scheme characteristic data corresponding to each initial scheme needs to be calculated and generated synchronously based on the personnel allocation results, power operation task data and historical shift data of each initial scheme. This includes personnel characteristics (skill matching degree, task work experience, regional familiarity, time matching degree, workload and performance score) and shift characteristics (shift experience matching degree and personnel utilization efficiency). This provides complete data input for the expert preference discrimination model to calculate the corresponding expert preference probability. It should be noted that the specific calculation process of the scheme characteristic data corresponding to the initial scheme can be referred to the relevant description of case feature extraction in the case feature dataset construction step above, and will not be repeated here.

[0039] Based on the normalized time cost and expert preference probability of each initial scheme, the corresponding comprehensive fitness value is calculated using the dual-drive fitness function, and multiple parent schemes are obtained based on the comprehensive fitness value. In practical applications, the process of obtaining multiple parent schemes using a tournament selection mechanism based on the comprehensive fitness value of each initial scheme in the current scheduling scheme population includes: randomly selecting several initial scheme individuals from the scheduling scheme population to form a group, selecting the initial scheme individual with the highest fitness as the parent, and repeating this process until the required number of parent schemes are selected.

[0040] Each parent scheme is sequentially subjected to crossover and mutation operations to obtain multiple offspring schemes. Based on all the offspring schemes, the optimal offspring scheme is obtained. In practical applications, the process of generating the optimal offspring scheme by sequentially performing genetic operations on each pair of parent schemes includes: first, performing single-point crossover, exchanging gene segments of the two parent schemes at a randomly selected crossover point to generate two new scheme individuals; then, performing uniform mutation on the crossover individuals, randomly changing the values ​​at gene loci with a certain probability to obtain offspring schemes; after all offspring schemes are generated, their corresponding comprehensive fitness values ​​are calculated, and the scheme with the highest comprehensive fitness is selected as the optimal offspring scheme.

[0041] The scheduling scheme population is updated based on all the parent schemes and all the optimal child schemes. The optimal scheme of the updated scheduling scheme population is obtained, and when the updated scheduling scheme population meets the preset iteration termination condition, the optimal scheme is taken as the target scheduling scheme. In practical applications, the process of updating the scheduling scheme population based on all parent and child optimal schemes includes: merging the parent schemes and the child optimal schemes, sorting them from high to low according to their comprehensive fitness values, and selecting the top N schemes with the highest comprehensive fitness (N being the size of the scheduling scheme population) to form a new generation of scheduling scheme population. This retains excellent schemes from the parent schemes while introducing the child optimal schemes, maintaining the diversity of the scheduling scheme population. After the update, the current optimal scheme is found from the new scheduling scheme population. The above iterative process is repeated until the preset iteration termination condition (such as reaching the maximum number of iterations or the comprehensive fitness showing no significant improvement for multiple generations) is met. At this point, the optimal scheme in the current scheduling scheme population is output as the final target scheduling scheme, thus obtaining a scheduling scheme that achieves optimal time cost and aligns with high-quality scheduling experience while meeting business constraints.

[0042] It should be noted that the iterative optimization implementation details of the personnel allocation optimization using the cultural gene algorithm in this embodiment can be found in existing technologies and will not be elaborated here. Furthermore, considering the potential conflicts between different maintenance tasks during actual scheduling due to overlapping time windows and duplicate personnel allocation, this embodiment integrates a local search strategy into the cultural gene algorithm to overcome the limitations of traditional genetic algorithms, which often struggle to effectively handle such detailed and specific constraint conflicts due to global search and are prone to getting stuck in local optima or generating numerous infeasible solutions under complex constraints. After identifying the aforementioned conflict, the cultural gene algorithm does not simply discard the scheduling plan or impose penalties, but instead initiates a local, heuristic adjustment process, adding a search for resolving temporal conflicts. In practical applications, for time conflicts between tasks, the execution time windows of the tasks are proactively adjusted to eliminate scheduling conflicts while satisfying task timing and dependencies. For example, we can first traverse the scheduling data matrix and extract the task sequence of each worker according to the time axis to construct a task linked list; traverse this linked list and compare the time attribute variables of adjacent tasks. If it is determined that adjacent tasks have time overlap (i.e., the expected completion time of the preceding task is later than the planned start time of the following task), it is determined to be a conflict and the corresponding task node identifier is recorded in the time conflict set; then, we start the time sequence conflict resolution, giving priority to the time shifting strategy of postponing the start time of the following task, under the premise of not violating the hard constraints such as the specified deadline and daily working hours (i.e., calculating the overlapping time difference and adding the following task accordingly). If the shift results in the task being overdue (i.e., the updated completion time is greater than the deadline), the original assignment of the task is revoked. Based on the relevant personnel skill status database, among currently available candidates who meet the skill requirements, the person with the lowest accumulated working hours (lowest workload) is prioritized to take over the task. Finally, a closed-loop check is performed, updating the time conflict set after each adjustment until it is empty to ensure the solution returns to the feasible solution space. If the conflict cannot be resolved after exhaustively trying all adjustment strategies, the local search is terminated and a fitness penalty is applied to the individual solution. This embodiment, while inheriting the advantages of global search in genetic algorithms, innovatively integrates machine learning-assisted decision-making mechanisms, local refined search strategies, and local conflict resolution mechanisms. This not only improves search efficiency but also enhances the applicability and executability of the scheduling scheme in real-world scenarios.

[0043] The present invention provides the following: After acquiring power operation task data, scheduling constraint rules, historical scheduling data including a success case library, an expert experience library, and a problem case library, and personnel skill data, an expert preference discrimination model is constructed based on the case feature dataset obtained by feature extraction from the historical scheduling data. Based on the power operation task data, the task time period is decomposed to obtain the target shift combination with the optimization objective of minimizing the total task operation cost. Based on the target shift combination, scheduling constraint rules, personnel skill data, and expert preference discrimination model, personnel allocation optimization is performed based on the cultural gene algorithm to obtain the power system operation personnel scheduling optimization scheme of the target scheduling scheme. Based on the two-stage optimization mechanism of first decomposing the task time period and then allocating personnel, it can not only ensure the quality and efficiency of scheduling optimization, but also take into account the feasibility of the scheduling scheme and the efficiency of personnel allocation, improve the practicality and reliability of the scheduling scheme, and can also achieve rapid response in the case of changes in business rules or changes in personnel skill structure without large-scale model parameter adjustments, effectively meet the application needs of different actual scheduling scenarios, and reduce the cost of use.

[0044] Furthermore, in order to provide relevant work scheduling decision-makers with an intuitive and reliable basis for decision-making, this embodiment preferably converts the target scheduling scheme obtained by the aforementioned personnel allocation optimization based on the cultural gene algorithm into structured information that can be directly referenced and used by managers; specifically, the method further includes: A target scheduling table is generated based on the target scheduling plan, and a comprehensive evaluation of the target scheduling table is conducted based on a preset indicator system to obtain the corresponding scheduling evaluation result. The target scheduling table can be understood as a specific executable plan generated based on the target scheduling plan, which may include information such as the daily work tasks of each operator, the maintenance schedule of each substation, the allocation of skill resources, and time path planning. In practical applications, the basic mapping relationship between worker IDs, task IDs, and times is first extracted based on the target scheduling scheme. Then, by using worker IDs and substation IDs as foreign keys, the relevant work information database is linked to complete the detailed information such as worker names and substation coordinates. Next, using worker IDs as the primary key, the daily assigned maintenance task nodes are arranged in ascending order by timestamp. Combined with the substation geographical coordinates, the Geographic Information System (GIS) interface is called or a preset inter-substation travel time matrix is ​​queried to calculate and insert the travel time between adjacent task nodes, forming a coherent path plan. Subsequently, according to the maintenance task type, the corresponding material database is searched to automatically match the required tools, materials, and skill qualification requirements, and these are bound as attribute fields to the corresponding task node data. Finally, a structured schedule is generated, and the above multi-dimensional reorganized data is filled and output according to a preset format, ultimately generating a structured target schedule for direct on-site dispatch.

[0045] In this embodiment, the preset indicator system can be understood as a quality indicator system for multi-dimensional comprehensive analysis and evaluation of the target scheduling schedule. To ensure the comprehensiveness and reliability of the scheduling quality evaluation, this embodiment preferably sets the preset indicator system to include task completion rate, resource utilization rate, skill matching degree, cost-effectiveness, and similarity of high-quality scheduling. Among them, the task completion rate is calculated by the ratio of the number of successfully arranged tasks in the target scheduling plan to the total number of required tasks. The resource utilization rate is the ratio of the total working hours actually planned to be allocated to the maximum available total working hours in the period. The skill matching degree is the average ratio of the actual skill level of the personnel allocated to each task node to the required skill level. The cost-effectiveness is the ratio of the preset benchmark number of dispatched workers (such as the average number of dispatched workers under the same task scale in history) to the total number of dispatched workers in the target scheduling plan.

[0046] In this embodiment, the similarity of high-quality scheduling is the average similarity between the target scheduling scheme and historical high-quality scheduling schemes, which can be expressed as: in, For workers in historically high-quality scheduling schemes The set of work hours and task types allows for the extraction of workers by retrieving all historical high-quality scheduling plans. This is obtained by combining the working time windows and specific task type information from various historical high-quality scheduling schemes. For the workers in the current target shift schedule The set of work hours and task types is used to extract the corresponding workers by parsing the currently generated target work schedule. i The allocated time window is obtained by combining features with specific task type information; Total number of workers; The average similarity between the target scheduling scheme calculated based on the Jaccard similarity coefficient and the historical high-quality scheduling scheme is calculated by determining the degree of overlap (intersection over union ratio) of each worker in terms of work time distribution and task skill attributes in the historical high-quality scheduling scheme and the current target scheduling scheme, and then calculating the average of all workers. This provides an intuitive reference for the acceptability and rationality of the target scheduling scheme.

[0047] By summarizing the target schedule obtained through the above methods and steps, as well as the scheduling evaluation results including task completion rate, resource utilization rate, skill matching degree, cost-effectiveness, and similarity of high-quality schedules, we can provide intuitive, reliable, and comprehensive decision-making basis for relevant work scheduling decision-makers.

[0048] This invention provides a technical solution that constructs an expert preference discrimination model based on a case feature dataset obtained by feature extraction from historical scheduling data. Based on power operation task data, it decomposes task time periods to obtain target shift combinations with the optimization objective of minimizing the total cost of task operations. Then, based on the target shift combinations, scheduling constraints, personnel skill data, and the expert preference discrimination model, it optimizes personnel allocation using a cultural gene algorithm to obtain a power system operation personnel scheduling optimization scheme. Finally, it generates a target scheduling table based on the target scheduling scheme and comprehensively evaluates the target scheduling table based on a preset indicator system to obtain the scheduling evaluation result. This two-stage optimization mechanism, based on task time period decomposition followed by personnel allocation, ensures the quality and efficiency of scheduling optimization, balances the feasibility of the scheduling scheme with the efficiency of personnel allocation, improves the practicality and reliability of the scheduling scheme, and enables rapid response to scenarios involving changes in business rules or personnel skill structures. It meets the application needs of different actual scheduling scenarios, reduces usage costs, and converts the target scheduling scheme into structured information that can be directly referenced by managers, providing intuitive and reliable decision-making basis for relevant operation scheduling decision-makers.

[0049] It should be noted that although the steps in the flowchart above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise explicitly stated in this document, there is no strict order requirement for the execution of these steps, and they can be executed in other orders.

[0050] In one embodiment, such as Figure 2 As shown, a power system operator scheduling optimization system is provided, the system comprising: Data acquisition module 1 is used to acquire power operation task data, scheduling constraint rules, historical scheduling data, and personnel skill data; the historical scheduling data includes a success case library, an expert experience library, and a problem case library; Model building module 2 is used to extract features from the historical scheduling data to obtain the corresponding case feature dataset, and to build an expert preference discrimination model based on the case feature dataset; Task decomposition module 3 is used to decompose the task time period based on the power operation task data, with the optimization objective of minimizing the total cost of the task operation, to obtain the target shift combination; The personnel allocation module 4 is used to optimize personnel allocation based on the cultural gene algorithm according to the target shift combination, the shift scheduling constraint rules, the personnel skill data and the expert preference discrimination model, so as to obtain the target shift scheduling scheme.

[0051] In one embodiment, a power system operator scheduling optimization system is provided, the system further comprising: The scheduling scheme evaluation module is used to generate a target scheduling table based on the target scheduling scheme, and to comprehensively evaluate the target scheduling table based on a preset indicator system to obtain the corresponding scheduling evaluation results; the preset indicator system includes task completion rate, resource utilization rate, skill matching degree, cost-effectiveness and similarity of high-quality scheduling.

[0052] Specific limitations regarding the power system operator scheduling optimization system can be found in the limitations of the power system operator scheduling optimization method described above; the corresponding technical effects are equivalent and will not be repeated here. Each module in the aforementioned power system operator scheduling optimization system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the corresponding operations of each module.

[0053] Figure 3 An internal structural diagram of a computer device is shown in one embodiment. This computer device may specifically be a terminal or a server. Figure 3 As shown, the computer device includes a processor, memory, network interface, display, camera, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When executed by the processor, the computer program can implement a method for optimizing the scheduling of power system workers. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0054] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. Specific computing devices may include more or fewer components than those shown in the figure, or combine certain components, or have the same component arrangement.

[0055] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described above.

[0056] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0057] In summary, the power system operator scheduling optimization method, system, computer equipment, and storage medium provided by this invention, based on a two-stage optimization mechanism of first decomposing task time periods and then allocating personnel, not only ensures the quality and efficiency of scheduling optimization, but also takes into account the feasibility of the scheduling scheme and the efficiency of personnel allocation, improving the practicality and reliability of the scheduling scheme. Furthermore, it can achieve rapid response to scenarios of changes in business rules or personnel skill structures without the need for large-scale model parameter adjustments, effectively meeting the application needs of different actual scheduling scenarios and reducing usage costs.

[0058] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. 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 in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0059] The above-described embodiments are merely preferred embodiments of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention. Therefore, the scope of protection of this invention should be determined by the scope of the claims.

Claims

1. A method for optimizing the scheduling of power system workers, characterized in that, The method includes: Acquire power operation task data, scheduling constraint rules, historical scheduling data, and personnel skill data; the historical scheduling data includes a success case database, an expert experience database, and a problem case database. Feature extraction is performed on the historical scheduling data to obtain the corresponding case feature dataset, and an expert preference discrimination model is constructed based on the case feature dataset; Based on the power operation task data, the task time period is decomposed with the goal of minimizing the total cost of the task operation, and the target shift combination is obtained. Based on the target shift combination, the shift scheduling constraints, the personnel skill data, and the expert preference discrimination model, personnel allocation is optimized using the cultural gene algorithm to obtain the target shift scheduling scheme.

2. The method for optimizing the scheduling of power system operators as described in claim 1, characterized in that, The case feature dataset includes case features and corresponding scheduling effect labels for each scheduling case; the case features include personnel features and scheduling features; the personnel features include the skill matching degree, task work experience, regional familiarity, time matching degree, workload and performance score of each operator; the scheduling features include scheduling experience matching degree and personnel utilization efficiency; the scheduling effect labels include high-quality scheduling labels and low-quality scheduling labels.

3. The method for optimizing the scheduling of power system operators as described in claim 1, characterized in that, The step of constructing an expert preference discrimination model based on the case feature dataset includes: The expert preference discrimination model is constructed by training the decision tree algorithm using the case features corresponding to the high-quality scheduling label and the low-quality scheduling label in the case feature dataset as positive and negative samples, respectively.

4. The method for optimizing the scheduling of power system operators as described in claim 1, characterized in that, The step of decomposing the task time period based on the power operation task data, with the optimization objective of minimizing the total task operation cost, to obtain the target shift combination includes: With minimizing the total cost of the task as the optimization objective, a task decomposition optimization function is constructed; Based on the task decomposition optimization function, a task decomposition optimization model is constructed based on preset optimization constraints; the preset optimization constraints include daily working hours constraints and total equipment maintenance requirements constraints. Based on the daily working hour constraints and the power operation task data, shifts are enumerated to obtain a candidate shift set; Based on the candidate shift set, the task decomposition optimization model is equivalently transformed to obtain the shift combination optimization model; The target train schedule combination is obtained by solving the optimization model of the train schedule combination.

5. The method for optimizing the scheduling of power system operators as described in claim 1, characterized in that, The step of optimizing personnel allocation based on the cultural gene algorithm according to the target shift combination, the shift scheduling constraint rules, the personnel skill data, and the expert preference discrimination model to obtain the target shift scheduling scheme includes: Based on the target shift combination, the scheduling constraint rules, and the personnel skill data, initialize the scheduling scheme population using heuristic rules; Based on the scheduling scheme population and the dual-drive fitness function, personnel allocation is optimized using the cultural gene algorithm to obtain the target scheduling scheme; the dual-drive fitness function is obtained by weighted fusion of the expert preference probability and normalized time cost of the scheduling scheme.

6. The method for optimizing the scheduling of power system operators as described in claim 5, characterized in that, The step of optimizing personnel allocation based on the cultural gene algorithm according to the scheduling scheme population and the preset dual-drive fitness function to obtain the target scheduling scheme includes: Based on the effective working time and total task delay of each initial scheme in the scheduling scheme population, a total time cost analysis is performed to obtain the corresponding normalized time cost. The scheme characteristics of each initial scheme are analyzed according to the expert preference discrimination model to obtain the corresponding expert preference probability; Based on the normalized time cost and expert preference probability of each initial scheme, the corresponding comprehensive fitness value is calculated based on the dual-drive fitness function, and multiple parent schemes are obtained based on the comprehensive fitness value. Each parent scheme is subjected to crossover and mutation operations in turn to obtain multiple child schemes, and the optimal child scheme is obtained based on all the child schemes. The scheduling scheme population is updated based on all the parent schemes and all the optimal child schemes. The optimal scheme of the updated scheduling scheme population is obtained, and when the updated scheduling scheme population meets the preset iteration termination condition, the optimal scheme is taken as the target scheduling scheme.

7. The method for optimizing the scheduling of power system operators as described in claim 1, characterized in that, The method further includes: A target scheduling table is generated based on the target scheduling scheme, and a comprehensive evaluation of the target scheduling table is conducted based on a preset indicator system to obtain the corresponding scheduling evaluation results; the preset indicator system includes task completion rate, resource utilization rate, skill matching degree, cost-effectiveness and similarity of high-quality scheduling.

8. A power system operator scheduling optimization system, characterized in that, The system includes: The data acquisition module is used to acquire power operation task data, scheduling constraint rules, historical scheduling data, and personnel skill data; the historical scheduling data includes a success case library, an expert experience library, and a problem case library. The model building module is used to extract features from the historical scheduling data to obtain the corresponding case feature dataset, and to build an expert preference discrimination model based on the case feature dataset. The task decomposition module is used to decompose the task time period based on the power operation task data, with the optimization objective of minimizing the total cost of the task operation, to obtain the target shift combination. The personnel allocation module is used to optimize personnel allocation based on the cultural gene algorithm according to the target shift combination, the scheduling constraint rules, the personnel skill data and the expert preference discrimination model, so as to obtain the target shift plan.

9. A computer 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 computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images