An intelligent system operation and maintenance method based on genetic algorithm

Abstract

The present invention provides an intelligent system operation and maintenance method based on a genetic algorithm, comprising the following steps: S101, obtaining all unprocessed operation and maintenance work orders of the intelligent system, and generating a first operation and maintenance work order set based on the obtained operation and maintenance work orders; S102, obtaining multiple operation and maintenance solutions corresponding to the operation and maintenance work orders in the first operation and maintenance work order set, and generating a first operation and maintenance solution set based on the obtained multiple operation and maintenance solutions; S103, inputting the first operation and maintenance work order set and the first operation and maintenance solution set into a genetic algorithm to obtain an optimal individual output by the genetic algorithm, wherein the goal of the genetic algorithm is to minimize the total cost of processing all operation and maintenance work orders; S104, executing the complete operation and maintenance solution represented by the optimal individual as the optimal operation and maintenance solution, wherein the complete operation and maintenance solution includes an optimal operation and maintenance work order set and an optimal operation and maintenance solution set, and the multiple operation and maintenance work orders in the optimal operation and maintenance work order set are arranged in the optimal solution order. The present invention can assist operation and maintenance personnel in quickly formulating reasonable operation and maintenance work order solutions and execution plans.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent system operation and maintenance, and in particular to an intelligent system operation and maintenance method based on a genetic algorithm. Background Art

[0002] With the development of information technology, it has become commonplace for enterprises to use various intelligent business management systems to process relevant data in their daily operations and management. The application of intelligent systems plays a crucial role in the operation and development of enterprises, so ensuring the smooth operation of these intelligent systems is essential. During the daily operation of intelligent systems, system users may encounter various problems. To address these problems, users can initiate maintenance tickets, record and upload the problems they encounter. The maintenance department then receives the tickets and provides maintenance solutions based on the user's description. For large-scale business systems, the number of maintenance tickets increases accordingly. When the maintenance department is unable to handle them in time, maintenance tickets tend to accumulate. In this case, it is obviously unreasonable to strictly handle maintenance tickets according to the order in which they were submitted. This is because the importance and urgency of maintenance tickets submitted earlier may not be as high as those submitted later, and the difficulty of resolving them may not be the same. Currently, maintenance departments handle maintenance tickets based on the subjective judgment of maintenance personnel. This can lead to urgent and important issues not being handled in a timely manner, resulting in system malfunctions and difficulties in the progress of related online business. Therefore, how to rationally prioritize the resolution of maintenance tickets is an urgent problem that needs to be solved. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide an intelligent system operation and maintenance method based on genetic algorithm, which reasonably arranges the processing order of operation and maintenance work orders of the intelligent system through genetic algorithm, thereby ensuring that key system problems can be handled in a timely manner on the basis of orderly processing of operation and maintenance work orders.

[0004] To achieve the above-mentioned object of the invention, the present invention provides an intelligent system operation and maintenance method based on genetic algorithm, which comprises the following steps:

[0005] S101. Acquire all unprocessed operation and maintenance work orders of the intelligent system, and generate a first operation and maintenance work order set based on the acquired unprocessed operation and maintenance work orders;

[0006] S102: Acquire multiple operation and maintenance solutions corresponding to the operation and maintenance work orders in the first operation and maintenance work order set, and generate a first operation and maintenance solution set based on the acquired multiple operation and maintenance solutions;

[0007] S103: Input the first set of operation and maintenance work orders and the first set of operation and maintenance solutions into a genetic algorithm to obtain an optimal individual output by the genetic algorithm. The goal of the genetic algorithm is to minimize the total cost of processing all operation and maintenance work orders.

[0008] S104. The complete operation and maintenance plan represented by the optimal individual is executed as the optimal operation and maintenance plan. The complete operation and maintenance plan includes an optimal operation and maintenance work order set and an optimal operation and maintenance plan set. The multiple operation and maintenance work orders in the optimal operation and maintenance work order set are arranged in the order of optimal solution. The arrangement order of the operation and maintenance plans in the optimal operation and maintenance plan set is the same as the arrangement order of their corresponding operation and maintenance work orders in the optimal operation and maintenance work orders.

[0009] Furthermore, in step S103, the genetic algorithm outputs the optimal individual based on the input first operation and maintenance work order set and the first operation and maintenance solution set, which specifically includes the following steps:

[0010] S201: Generate multiple individuals based on a first set of operation and maintenance work orders and a first set of operation and maintenance solutions to form an initial population;

[0011] S202, selecting individuals from the initial population using a particle swarm algorithm and outputting the selected individuals;

[0012] S203, performing a crossover operation and / or a mutation operation on the individuals output in the previous step;

[0013] S204. Iterate steps S202 and S203 until the algorithm converges and outputs the population.

[0014] Furthermore, step S201 specifically includes the following steps:

[0015] S301. Randomly shuffle the order of the operation and maintenance work orders in the first operation and maintenance work order set to form a new order of the operation and maintenance work orders as the second operation and maintenance work order set;

[0016] S302. The operation and maintenance work orders in the reordered operation and maintenance work order set are polled in order, and at least one operation and maintenance solution corresponding to each operation and maintenance work order is found in order. One of the operation and maintenance solutions is randomly selected. After the operation and maintenance work orders are polled, all the selected operation and maintenance solutions are sorted in the same order as the operation and maintenance work orders to form a second set of operation and maintenance solutions.

[0017] S303: Combine the second operation and maintenance work order set and the second operation and maintenance plan set into a complete operation and maintenance plan, which is recorded as an individual;

[0018] S304: Count whether the number of individuals generated meets the initial population number requirement. If so, execute step S202; otherwise, execute step S301.

[0019] Furthermore, in step S102, multiple operation and maintenance solutions corresponding to the operation and maintenance work orders in the first operation and maintenance work order set are obtained, which specifically includes the following steps:

[0020] S401. Create a forward-order task QtoS and a reverse-order task StoQ. QtoS and StoQ each achieve their task objectives using a pre-trained model. The task objective of QtoS is to generate an operation and maintenance solution that can be used to solve the operation and maintenance work order problem based on the operation and maintenance work order. The task objective of StoQ is to generate possible operation and maintenance work orders based on existing operation and maintenance solutions, and record the possible operation and maintenance work orders as simulated work orders.

[0021] S402. Perform alternating iterative training on QtoS and StoQ to continuously generate new simulated work orders and operation and maintenance plans. After reaching the training target, output the simulated work orders and operation and maintenance plans generated in step S402.

[0022] S403: Match the simulated work order with the operation and maintenance work orders in the first operation and maintenance work order set, and select the successfully matched simulated work orders;

[0023] S404: Query the operation and maintenance plan generated in step S402 and corresponding to the screened simulated work order, and output it.

[0024] Furthermore, in step S401, after creating QtoS and StoQ, paired work order-solution datasets are used to pre-train the task models of QtoS and StoQ, respectively, to obtain corresponding pre-trained models.

[0025] Furthermore, in step S402, QtoS and StoQ are alternately iterated and trained, specifically including the following steps:

[0026] S501. During the first iterative training, the QtoS pre-trained model is first used to generate a corresponding operation and maintenance plan for the operation and maintenance work order training set. The operation and maintenance plan generated by QtoS is then input into StoQ to generate a simulated work order. The pre-trained model parameters of QtoS and StoQ are updated to obtain the updated pre-trained model, completing the first iterative training.

[0027] S502. In subsequent iterative training processes, the training data sets used each time are the new simulated work orders and operation and maintenance plan data generated by the pre-training models of QtoS and StoQ in the previous iteration, respectively, until the pre-training model reaches convergence and the iteration is stopped.

[0028] Furthermore, in step S203, the mutation operation includes the following types:

[0029] Randomly select an operation and maintenance work order from the second operation and maintenance work order set and change its corresponding operation and maintenance plan;

[0030] Randomly select at least two operation and maintenance work orders from the second operation and maintenance work order combination, change the relative order of the selected operation and maintenance work orders, and adjust the order of their corresponding operation and maintenance solutions accordingly;

[0031] An operation and maintenance plan is randomly selected from the second operation and maintenance plan set, at least two corresponding operation and maintenance work orders are found from the second operation and maintenance work order set, and the relative order of the found operation and maintenance work orders is adjusted.

[0032] Furthermore, in step S203, the crossover operation includes the following types:

[0033] First crossover type: Randomly select two individuals from the population as the parent and mother. Split the second operation and maintenance work order sets of the parent and mother into two parts respectively. Concatenate the second operation and maintenance work order sets of the parent and mother to form a new operation and maintenance work order set. Perform the same operation on the second operation and maintenance plan sets of the parent and mother to form a new operation and maintenance plan set. Then, form a new individual based on the new operation and maintenance work order set and the new operation and maintenance plan set.

[0034] Second crossover type: Randomly select two individuals from the population as the father and mother, split the second operation and maintenance work order set of the father and mother into three parts respectively, splice the second operation and maintenance work order set after the father is split and the second operation and maintenance work order set after the mother is split to form a new operation and maintenance work order set, perform the same operation on the second operation and maintenance plan set of the father and mother to form a new operation and maintenance plan set, and form a new individual based on the new operation and maintenance work order set and operation and maintenance plan set.

[0035] Furthermore, in step S203, a crossover operation and / or a mutation operation is performed on the individual output in the previous step. Specifically, each time a mutation / crossover operation is performed, a type of mutation / crossover operation is randomly selected for execution.

[0036] Furthermore, the total cost of the operation and maintenance work order in step S103 is the sum of resource cost and delay cost, and the resource cost includes computing resource cost and human resource cost.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The present invention provides an intelligent system operation and maintenance method based on a genetic algorithm. The method generates a first operation and maintenance work order set based on unprocessed system operation and maintenance work orders, obtains a first operation and maintenance plan set based on the first operation and maintenance work order set, and inputs the first operation and maintenance work order set and the first operation and maintenance plan set into the genetic algorithm to obtain the optimal individual obtained by the genetic algorithm through calculation and processing with the goal of minimizing the total cost of the operation and maintenance work order. The complete operation and maintenance plan represented by the optimal individual is executed as the optimal operation and maintenance plan. The method can assist operation and maintenance personnel in quickly formulating reasonable operation and maintenance work order solutions and execution plans when the number of operation and maintenance work orders is large, thereby ensuring that important and urgent matters can be handled first, improving operation and maintenance efficiency, and better guaranteeing the smooth operation of various intelligent business systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only preferred embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0040] Figure 1 This is a schematic diagram of the overall process of an intelligent system operation and maintenance method based on genetic algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The principles and features of the present invention are described below with reference to the accompanying drawings. The enumerated embodiments are only used to explain the present invention and are not used to limit the scope of the present invention.

[0042] Reference Figure 1 This embodiment provides an intelligent system operation and maintenance method based on a genetic algorithm, the method comprising the following steps:

[0043] S101: Acquire all unprocessed operation and maintenance work orders of the intelligent system, and generate a first operation and maintenance work order set based on the acquired unprocessed operation and maintenance work orders.

[0044] S102: Acquire multiple corresponding operation and maintenance solutions from the operation and maintenance work orders in the first operation and maintenance work order set, and generate a first operation and maintenance solution set based on the acquired multiple operation and maintenance solutions. It is understandable that for the same operation and maintenance work order, there may be multiple operation and maintenance solutions that can be used to solve the problem described therein.

[0045] S103. Input the first set of operation and maintenance work orders and the first set of operation and maintenance solutions into a genetic algorithm to obtain the optimal individual output by the genetic algorithm. The goal of the genetic algorithm is to minimize the total cost of processing all operation and maintenance work orders. In this embodiment, the total cost of an operation and maintenance work order is the sum of resource costs and delay costs. Resource costs include computing resource costs and human resource costs. The maximum computing resource costs and human resource costs cannot exceed the upper limit of the operation and maintenance department's available computing resources and human resources. Delay costs are the costs incurred by the enterprise when the problem described in the operation and maintenance work order remains unresolved.

[0046] S104. Execute the complete operation and maintenance plan represented by the optimal individual as the optimal operation and maintenance plan. The complete operation and maintenance plan includes a set of optimal operation and maintenance work orders and a set of optimal operation and maintenance plans. The multiple operation and maintenance work orders in the set of optimal operation and maintenance work orders are arranged in the order of optimal solution. The order of the operation and maintenance plans in the set of optimal operation and maintenance plans is the same as the order of their corresponding operation and maintenance work orders in the set of optimal operation and maintenance plans. In the optimal operation and maintenance plan, each operation and maintenance work order corresponds to an optimal operation and maintenance plan.

[0047] The optimal order for resolving operation and maintenance work orders follows the following principles: operation and maintenance work orders with high delay costs are placed first, and operation and maintenance work orders with low delay costs are placed last; when the delay costs are the same, if the operation and maintenance department has limited computing resources and human resources, operation and maintenance work orders with high resource costs should be placed first, and operation and maintenance work orders with low resource costs should be placed last, because operation and maintenance work orders with high resource costs often require more resources and time to resolve. If they cannot be resolved in time, they may have a greater negative impact on the enterprise; if the operation and maintenance department has sufficient computing resources and human resources, operation and maintenance work orders with low resource costs can be placed first, and operation and maintenance work orders with high resource costs can be placed last, giving priority to resolving operation and maintenance work orders that consume less resources. This way, more operation and maintenance work orders can be resolved faster, thereby improving work efficiency and benefits.

[0048] In step S103, the genetic algorithm outputs the optimal individual based on the input first operation and maintenance work order set and the first operation and maintenance solution set, which specifically includes the following steps:

[0049] S201. Generate multiple individuals based on a first operation and maintenance work order set and a first operation and maintenance plan set to form an initial population.

[0050] S202: Select individuals from the initial population using a particle swarm algorithm and output the selected individuals.

[0051] S203: Perform a crossover operation and / or a mutation operation on the individuals output in the previous step.

[0052] S204. Iterate steps S202 and S203 until the algorithm converges and outputs the population.

[0053] As a preferred example, generating multiple individuals based on the first operation and maintenance work order set and the first operation and maintenance plan set to form an initial population specifically includes the following steps:

[0054] S301. Randomly shuffle the order of the operation and maintenance work orders in the first operation and maintenance work order set to form a new order of the operation and maintenance work orders as the second operation and maintenance work order set.

[0055] S302. The operation and maintenance work orders in the reordered operation and maintenance work order set are sequentially polled, and at least one operation and maintenance solution corresponding to each operation and maintenance work order is found in sequence. One of the operation and maintenance solutions is randomly selected. After the operation and maintenance work orders are polled, all selected operation and maintenance solutions are sorted in the same order as the corresponding operation and maintenance work orders to form a second set of operation and maintenance solutions.

[0056] S303: Combine the second operation and maintenance work order set and the second operation and maintenance plan set into a complete operation and maintenance plan, which is recorded as an individual.

[0057] S304: Count whether the number of individuals generated meets the initial population number requirement. If so, execute step S202; otherwise, execute step S301.

[0058] As a preferred example, in step S102, obtaining multiple operation and maintenance solutions corresponding to the operation and maintenance work orders in the first operation and maintenance work order set specifically includes the following steps:

[0059] S401. Create a forward-order task QtoS and a reverse-order task StoQ. After creating QtoS and StoQ, pre-train the task models for QtoS and StoQ using paired work order-solution datasets, thereby obtaining corresponding pre-trained models. QtoS and StoQ each achieve their task objectives through the pre-trained models. The task objective of QtoS is to generate an operation and maintenance solution that can be used to resolve the operation and maintenance work order problem based on the operation and maintenance work order. The task objective of StoQ is to generate possible operation and maintenance work orders based on existing operation and maintenance solutions. Possible operation and maintenance work orders are recorded as simulated work orders.

[0060] S402. Perform alternating iterative training on QtoS and StoQ to continuously generate new simulated work orders and operation and maintenance plans. After reaching the training goal, output the simulated work orders and operation and maintenance plans generated in step S402.

[0061] S403: Match the simulated work order with the operation and maintenance work orders in the first operation and maintenance work order set, and select the successfully matched simulated work orders.

[0062] S404: Query the operation and maintenance plan generated in step S402 and corresponding to the screened simulated work order, and output it.

[0063] Among them, QtoS and StoQ are alternately iterated and trained, specifically including the following steps:

[0064] S501. During the first iterative training, the pre-trained model of QtoS is first used to generate a corresponding operation and maintenance plan for the operation and maintenance work order training set. The operation and maintenance plan generated by QtoS is then input into StoQ to generate a simulated work order. The pre-trained model parameters of QtoS and StoQ are updated to obtain the updated pre-trained model and complete the first iterative training.

[0065] S502. In subsequent iterative training processes, the training data sets used each time are the new simulated work orders and operation and maintenance plan data generated by the pre-training models of QtoS and StoQ in the previous iteration, respectively, until the pre-training model reaches convergence and the iteration is stopped.

[0066] The method provided in this embodiment first generates an operation and maintenance solution that can be used to solve an operation and maintenance work order through the forward-order task QtoS. Then, through the reverse-order task StoQ, the operation and maintenance solution generated by the forward-order task QtoS is used to infer possible simulated work orders. Based on the simulated work orders, more operation and maintenance solutions are inferred. Through continuous cyclic training, a large number of operation and maintenance solutions are generated, thereby increasing the number of individuals in the population. For the operation and maintenance solutions generated by the forward-order and reverse-order tasks, it is necessary to check whether they can be applied to the operation and maintenance work orders in the first operation and maintenance work order set, so as to screen out invalid solutions and improve the quality of the individuals in the population.

[0067] In this embodiment, the mutation operations performed on individuals may include the following types:

[0068] One is to randomly select an operation and maintenance work order from the second operation and maintenance work order set and change its corresponding operation and maintenance plan.

[0069] The second is to randomly select at least two operation and maintenance work orders from the second operation and maintenance work order combination, change the relative order of the selected operation and maintenance work orders, and adjust the order of their corresponding operation and maintenance plans accordingly.

[0070] The third is to randomly select an operation and maintenance plan from the second operation and maintenance plan set, find at least two corresponding operation and maintenance work orders from the second operation and maintenance work order set, and adjust the relative order of the found operation and maintenance work orders.

[0071] The crossover operations performed on individuals can include the following types:

[0072] The first crossover type: Randomly select two individuals from the population as the father and mother, split the second operation and maintenance work order set of the father and mother into two parts respectively, splice the second operation and maintenance work order set after the father is split and the second operation and maintenance work order set after the mother is split to form a new operation and maintenance work order set, perform the same operation on the second operation and maintenance plan set of the father and mother to form a new operation and maintenance plan set, and form a new individual based on the new operation and maintenance work order set and operation and maintenance plan set.

[0073] The second crossover type: Randomly select two individuals from the population as the parent and mother. Split the second operation and maintenance work order sets of the parent and mother into three parts respectively. Combine the second operation and maintenance work order sets of the parent and mother to form a new operation and maintenance work order set. Perform the same operation on the second operation and maintenance plan sets of the parent and mother to form a new operation and maintenance plan set. Based on this new operation and maintenance work order set and operation and maintenance plan set, a new individual is formed. It is important to note that regardless of the type of crossover operation, the correspondence between operation and maintenance work orders and operation and maintenance plans cannot be changed.

[0074] Obviously, according to different actual needs and application scenarios, the mutation operation and crossover operation performed on individuals may also include other types, which are not listed one by one in this embodiment.

[0075] At the same time, when performing crossover and / or mutation operations on individuals, one type of mutation operation or crossover operation is randomly selected for execution.

[0076] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent system operation and maintenance method based on genetic algorithm, characterized in that: The method comprises the following steps: S101. Acquire all unprocessed operation and maintenance work orders of the intelligent system, and generate a first operation and maintenance work order set based on the acquired unprocessed operation and maintenance work orders; S102: Acquire multiple operation and maintenance solutions corresponding to the operation and maintenance work orders in the first operation and maintenance work order set, and generate a first operation and maintenance solution set based on the acquired multiple operation and maintenance solutions; S103: Input the first set of operation and maintenance work orders and the first set of operation and maintenance solutions into a genetic algorithm to obtain an optimal individual output by the genetic algorithm. The goal of the genetic algorithm is to minimize the total cost of processing all operation and maintenance work orders. S104: Execute the complete operation and maintenance plan represented by the optimal individual as the optimal operation and maintenance plan. The complete operation and maintenance plan includes an optimal operation and maintenance work order set and an optimal operation and maintenance plan set. The multiple operation and maintenance work orders in the optimal operation and maintenance work order set are arranged in the order of optimal solution. The arrangement order of the operation and maintenance plans in the optimal operation and maintenance plan set is the same as the arrangement order of the corresponding operation and maintenance work orders in the optimal operation and maintenance work order set. In step S102, multiple operation and maintenance solutions corresponding to the operation and maintenance work orders in the first operation and maintenance work order set are obtained, which specifically includes the following steps: S401. Create a forward-order task QtoS and a reverse-order task StoQ. QtoS and StoQ each achieve their task objectives through a pre-trained model. The task objective of QtoS is to generate an operation and maintenance solution that can be used to solve the operation and maintenance work order problem based on the operation and maintenance work order. The task objective of StoQ is to generate possible operation and maintenance work orders based on existing operation and maintenance solutions. The possible operation and maintenance work orders are recorded as simulated work orders. After creating QtoS and StoQ, the task models of QtoS and StoQ are pre-trained using paired work order-solution datasets, respectively, to obtain corresponding pre-trained models. S402. Perform alternating iterative training on QtoS and StoQ to continuously generate new simulated work orders and operation and maintenance plans. After reaching the training target, output the simulated work orders and operation and maintenance plans generated in step S402. S403: Match the simulated work order with the operation and maintenance work orders in the first operation and maintenance work order set, and select the successfully matched simulated work orders; S404: Query the operation and maintenance plan generated in step S402 and corresponding to the simulated work order obtained by screening, and output it; In step S402, QtoS and StoQ are alternately iterated and trained, specifically including the following steps: S501. During the first iterative training, the QtoS pre-trained model is first used to generate a corresponding operation and maintenance plan for the operation and maintenance work order training set. The operation and maintenance plan generated by QtoS is then input into StoQ to generate a simulated work order. The pre-trained model parameters of QtoS and StoQ are updated to obtain the updated pre-trained model, completing the first iterative training. S502. In subsequent iterative training processes, the training data sets used each time are the new simulated work orders and operation and maintenance plan data generated by the pre-training models of QtoS and StoQ in the previous iteration, respectively, until the pre-training model reaches convergence and the iteration is stopped.

2. The intelligent system operation and maintenance method based on genetic algorithm according to claim 1, characterized in that: In step S103, the genetic algorithm outputs the optimal individual based on the input first operation and maintenance work order set and the first operation and maintenance solution set, which specifically includes the following steps: S201: Generate multiple individuals based on a first set of operation and maintenance work orders and a first set of operation and maintenance solutions to form an initial population; S202, selecting individuals from the initial population using a particle swarm algorithm and outputting the selected individuals; S203, performing a crossover operation and / or a mutation operation on the individuals output in the previous step; S204. Iterate steps S202 and S203 until the algorithm converges and outputs the population.

3. The intelligent system operation and maintenance method based on genetic algorithm according to claim 2, characterized in that: Step S201 specifically includes the following steps: S301. Randomly shuffle the order of the operation and maintenance work orders in the first operation and maintenance work order set to form a new order of the operation and maintenance work orders as the second operation and maintenance work order set; S302. The operation and maintenance work orders in the reordered operation and maintenance work order set are polled in order, and at least one operation and maintenance solution corresponding to each operation and maintenance work order is found in order. One of the operation and maintenance solutions is randomly selected. After the operation and maintenance work orders are polled, all the selected operation and maintenance solutions are sorted in the same order as the operation and maintenance work orders to form a second set of operation and maintenance solutions. S303: Combine the second operation and maintenance work order set and the second operation and maintenance plan set into a complete operation and maintenance plan, which is recorded as an individual; S304: Count whether the number of individuals generated meets the initial population number requirement. If so, execute step S202; otherwise, execute step S301.

4. The intelligent system operation and maintenance method based on genetic algorithm according to claim 2, characterized in that: In step S203, the mutation operation includes the following types: Randomly select an operation and maintenance work order from the second operation and maintenance work order set and change its corresponding operation and maintenance plan; Randomly select at least two operation and maintenance work orders from the second operation and maintenance work order combination, change the relative order of the selected operation and maintenance work orders, and adjust the order of their corresponding operation and maintenance solutions accordingly; An operation and maintenance plan is randomly selected from the second operation and maintenance plan set, at least two corresponding operation and maintenance work orders are found from the second operation and maintenance work order set, and the relative order of the found operation and maintenance work orders is adjusted.

5. The intelligent system operation and maintenance method based on genetic algorithm according to claim 4, characterized in that: In step S203, the crossover operation includes the following types: First crossover type: Randomly select two individuals from the population as the parent and mother. Split the second operation and maintenance work order sets of the parent and mother into two parts respectively. Concatenate the second operation and maintenance work order sets of the parent and mother to form a new operation and maintenance work order set. Perform the same operation on the second operation and maintenance plan sets of the parent and mother to form a new operation and maintenance plan set. Then, form a new individual based on the new operation and maintenance work order set and the new operation and maintenance plan set. Second crossover type: Randomly select two individuals from the population as the father and mother, split the second operation and maintenance work order set of the father and mother into three parts respectively, splice the second operation and maintenance work order set after the father is split and the second operation and maintenance work order set after the mother is split to form a new operation and maintenance work order set, perform the same operation on the second operation and maintenance plan set of the father and mother to form a new operation and maintenance plan set, and form a new individual based on the new operation and maintenance work order set and operation and maintenance plan set.

6. The intelligent system operation and maintenance method based on genetic algorithm according to claim 5, characterized in that: In step S203, a crossover operation and / or a mutation operation is performed on the individuals output in the previous step. Specifically, each time a mutation / crossover operation is performed, a type of mutation / crossover operation is randomly selected for execution.

7. The intelligent system operation and maintenance method based on genetic algorithm according to claim 1, characterized in that: The total cost of the operation and maintenance work order in step S103 is the sum of resource cost and delay cost, and the resource cost includes computing resource cost and human resource cost.

Images