Scheduling method, system, device and storage medium for mining equipment maintenance resources

By building and optimizing the directed graph and decision tree of maintenance resources of mining equipment, and dynamically adjusting the resource scheduling plan, the problems of low resource scheduling efficiency and high cost in the existing technology are solved, and more efficient resource utilization and stability are achieved.

CN118428697BActive Publication Date: 2024-11-08YANTAI JIERUI NETWORK TRADING
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410881469.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-03
Publication Date
2024-11-08
Estimated Expiration
2044-07-03

AI Technical Summary

Technical Problem

In the electronic data resource scheduling of mining equipment, the prior art is low efficiency, high cost, low resource utilization, and lacks adaptability and interpretability, so it is unable to effectively deal with the constraints and goals of dynamic changes.

Method used

By constructing a directed graph of maintenance resources, transforming it into a hierarchical general graph, arranging and combining the initial plan, performing decision tree processing, filtering and optimizing the plan, resource maintenance combined with time and location constraints, and dynamically adjusting the scheduling plan to improve efficiency and stability.

Benefits of technology

It reduces maintenance costs, shortens maintenance time, improves resource utilization, and improves scheduling stability and computing efficiency.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The present application relates to the technical field of resource allocation in electronic data processing, specifically to a scheduling method, system, device and storage medium for mine equipment maintenance resources; the method of the present application first constructs a hierarchical general graph of the same resource type at the same level based on a directed graph of maintenance resources, and arranges and combines resources to obtain several basic plans, deletes basic plans with conflicts between resource feasible domains, and obtains several initial plans; then, the state variables of each initial plan are discretized and arranged and combined in the form of a decision tree to obtain a decision system, and according to the matching degree of each sub-decision plan in the decision system with the task tree to be inspected, the most suitable task solution to solve the problem is screened out; finally, the maintenance process sequence is determined based on the task solution, and the mine equipment resource maintenance is carried out; the method can reduce the maintenance cost, shorten the maintenance time, and improve the resource utilization rate, and has good calculation stability.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of resource allocation in electronic data processing, and in particular to a scheduling method, system, device and storage medium for mining equipment maintenance resources. Background Art

[0002] Mining equipment is the core element of mining production, and the status of equipment directly affects mining production and safety benefits. Reasonable scheduling of various maintenance resources (personnel, tools, spare parts, etc.) and formulating scientific maintenance plans are prerequisites for completing high-quality maintenance operations.

[0003] When performing resource scheduling based on electronic data of mining equipment, existing technologies mainly solve the problem through the following solutions. The manual experience method has a certain degree of interpretability, and the decision-making process can be understood and controlled by people; but it is easy to over-rely on past experience. When the scale and complexity of the problem are large, it is often inefficient and of low quality, resulting in high cost and time consumption of the maintenance method and low resource utilization. The mathematical programming algorithm can accurately model constraints and objective functions, and the solution process is rigorous; but the solution efficiency is low, and the calculation cost is expensive when the instance scale is large. At the same time, it lacks adaptive capabilities and cannot handle dynamically changing constraints and goals, resulting in high cost and time consumption of the maintenance method. The machine learning algorithm can automatically model without manual feature extraction; but it requires a large amount of labeled data, lacks the ability to migrate to new problems, and also lacks interpretability and controllability, resulting in poor stability of the scheduling method. Summary of the invention

[0004] The object of the present invention is to provide a method, system, device and storage medium for scheduling mining equipment maintenance resources which can reduce costs, shorten maintenance time, improve resource utilization and resource scheduling stability.

[0005] The technical solution of the present invention is as follows:

[0006] A method for scheduling mining equipment maintenance resources includes the following operations:

[0007] S1. Based on the maintenance process, personnel, tools and spare parts, a maintenance resource directed graph is constructed; in the maintenance resource directed graph, the nodes are maintenance processes, personnel, tools and spare parts, and the edges are the dependencies between maintenance processes, personnel, tools and spare parts; according to the resource type, the maintenance resource directed graph is converted into a hierarchical general graph of the same resource type at the same level; in the hierarchical general graph, the resources at the highest level and with dependencies between each other are arranged and combined to obtain several root solutions; in the order of the hierarchical general graph from high to low, the resources with dependencies on the root solutions are arranged and combined with several root solutions respectively to obtain several basic solutions; the basic solutions with conflicts between the feasible domains of internal resources are deleted to obtain several initial solutions;

[0008] S2. After discretizing the state variables of each of the initial solutions, the state variables are arranged and combined in the form of a decision tree to obtain a decision system containing several sub-decision solutions; the task of mine equipment to be inspected is converted into a task tree to be inspected in the form of a decision tree; the sub-decision solution corresponding to the maximum matching degree of the task tree to be inspected in the decision system is used as the task solution;

[0009] S3. Determine the maintenance process sequence based on the maintenance process information in the task solution; perform maintenance of mining equipment resources according to the maintenance process sequence, corresponding personnel, tools and spare parts, and time constraints and location constraints in the mining equipment maintenance tasks.

[0010] Screening several initial solutions to obtain several optimized initial solutions, which are used to replace several initial solutions to perform the operations in S2; the operations of the screening process are specifically as follows: constructing an initial scoring function based on time consumption, initial time consumption weight, cost, initial cost weight, utilization rate and initial utilization rate weight; obtaining an initial scoring value of each initial solution based on the initial scoring function; taking an initial solution with an initial scoring value greater than an initial scoring value threshold as a first solution; all first solutions form a first solution set; obtaining a scoring variance of the first solution set; if the scoring variance is greater than a mean threshold, increasing the initial time consumption weight, and / or the initial cost weight, and / or the initial utilization rate weight to the first time consumption weight, and / or the first cost weight, and / or the first utilization rate weight, respectively; updating the initial scoring function based on the first time consumption weight, and / or the first cost weight, and / or the first utilization rate weight to obtain an updated scoring function; obtaining an updated scoring value of each initial solution based on the updated scoring function, and taking an initial solution with an updated scoring value greater than the updated scoring threshold as an optimized initial solution.

[0011] If the score variance is greater than the mean threshold, and the difference between the score mean of the first solution set and the expected value of time consumption, and / or cost, and / or utilization is respectively less than the time consumption difference threshold, and / or the cost difference threshold, and / or the utilization difference threshold; the initial time consumption weight, and / or the initial cost weight, and / or the initial utilization weight are respectively reduced to the second time consumption weight, and / or the second cost weight, and / or the second utilization weight; based on the second time consumption weight, and / or the second cost weight, and / or the second utilization weight, the initial scoring function is updated to obtain an updated scoring function; the updated scoring function is used to execute the operation of obtaining an optimized initial solution.

[0012] The operation of reducing the initial time consumption weight, or the initial cost weight, or the initial utilization weight is achieved through the following formula: w 2 = w 0 ·(1-α·(uE)) , w2 is the second weight of time consumption, or the second weight of cost, or the second weight of utilization rate, w 0 is the initial time consumption weight, or the initial cost weight, or the initial utilization weight, α is the adjustment coefficient, u is the mean score, E It is the expected value of time consumption, or the expected value of cost, or the expected value of utilization rate.

[0013] If there are several task solutions in S2, and the task of mine equipment to be repaired changes; based on the mapping relationship between the state variables before the change of the task of mine equipment to be repaired and the state variables after the change, adjust the corresponding variables in each task solution to obtain several task solution update solutions; obtain the process execution sequence of each task solution update solution, as well as the personnel, tools and spare parts, and perform encoding processing to obtain the process dependency gene coding and process resource gene coding of each task solution update solution, and splice the process dependency gene coding and process resource gene coding of each task solution update solution respectively to obtain the solution scheduling code of each task solution update solution; use the solution scheduling code with fitness greater than the fitness threshold and density less than the density threshold as the new task solution update solution to solve the changed mine equipment to be repaired task.

[0014] The fitness is the sum of the encoding values ​​of the solution scheduling encoding; the density is the mean distance between the task solution update solution and other task solution update solutions; the distance is the product of the solution scheduling encoding and the other solution scheduling encoding.

[0015] In S2, if the state variable of the task to be inspected for mining equipment is a specific data, the matching degree is the product of the feature vector of the task tree to be inspected and the feature vector of the sub-decision plan.

[0016] A system for executing a scheduling method for mining equipment maintenance resources, comprising:

[0017] Perception cells are used to construct a directed graph of maintenance resources based on maintenance procedures, personnel, tools and spare parts; in the directed graph of maintenance resources, the nodes are maintenance procedures, personnel, tools and spare parts, and the edges are the dependencies between maintenance procedures, personnel, tools and spare parts; according to the resource type, the directed graph of maintenance resources is converted into a hierarchical graph of the same resource type at the same level; in the hierarchical graph, the resources at the highest level that have dependencies on each other are arranged and combined to obtain several root solutions; in the order of the hierarchical graph from high to low, the resources that have dependencies on the root solutions are arranged and combined with several root solutions respectively to obtain several basic solutions; basic solutions with conflicts between the feasible domains of internal resources are deleted to obtain several initial solutions;

[0018] The decision cell is used to discretize the state variables of each of the initial solutions and arrange and combine them in the form of a decision tree to obtain a decision system containing several sub-decision solutions; convert the task of mine equipment to be repaired into a task tree to be inspected in the form of a decision tree; and take the sub-decision solution in the decision system corresponding to the maximum matching degree of the task tree to be inspected as the task solution;

[0019] The execution cell is used to determine the maintenance process sequence based on the maintenance process information in the task solution; according to the maintenance process sequence, the corresponding personnel, tools and spare parts, as well as the time constraints and location constraints in the mine equipment maintenance tasks, the mine equipment resources are maintained.

[0020] A scheduling device for mining equipment maintenance resources comprises a processor and a memory, wherein the processor implements the above-mentioned scheduling method for mining equipment maintenance resources when executing a computer program stored in the memory.

[0021] A computer-readable storage medium is used to store a computer program, wherein the computer program implements the above-mentioned method for scheduling mining equipment maintenance resources when executed by a processor.

[0022] The beneficial effects of the present invention are:

[0023] The present invention provides a scheduling method for mine equipment maintenance resources. First, based on the maintenance process, and personnel, tools and spare parts that are updated in real time and available, a maintenance resource directed graph that is convenient for clarifying the relationship between resources is constructed; then, based on the maintenance resource directed graph, a hierarchical general graph of the same resource type at the same level is constructed, and the resources are arranged and combined in order from high to low to obtain a number of basic schemes, and the basic schemes with conflicts between resource feasible domains are deleted to obtain a number of reasonable initial schemes; then, the state variables of each initial scheme are discretized and arranged and combined in the form of a decision tree to achieve refinement and expansion of each initial scheme, and a decision system composed of a number of sub-decision schemes in the form of decision trees with a clear structure and easy to analyze is obtained, and according to the matching degree between each sub-decision scheme in the decision system and the task tree to be inspected, the most suitable task solution for solving the problem is screened out; finally, the maintenance process sequence is determined based on the task solution, and the mine equipment resource maintenance is carried out; the method can reduce the maintenance cost, shorten the maintenance time, and improve the resource utilization rate, and has good calculation stability. DETAILED DESCRIPTION

[0024] This embodiment provides a method for scheduling mining equipment maintenance resources, including the following operations:

[0025] S1. Based on the maintenance process, personnel, tools and spare parts, a maintenance resource directed graph is constructed; in the maintenance resource directed graph, the nodes are maintenance processes, personnel, tools and spare parts, and the edges are the dependencies between maintenance processes, personnel, tools and spare parts; according to the resource type, the maintenance resource directed graph is converted into a hierarchical general graph of the same resource type at the same level; in the hierarchical general graph, the resources at the highest level and with dependencies between each other are arranged and combined to obtain several root solutions; in the order of the hierarchical general graph from high to low, the resources with dependencies on the root solutions are arranged and combined with several root solutions respectively to obtain several basic solutions; the basic solutions with conflicts between the feasible domains of internal resources are deleted to obtain several initial solutions;

[0026] S2. After discretizing the state variables of each of the initial solutions, the state variables are arranged and combined in the form of a decision tree to obtain a decision system containing several sub-decision solutions; the task of mine equipment to be inspected is converted into a task tree to be inspected in the form of a decision tree; the sub-decision solution corresponding to the maximum matching degree of the task tree to be inspected in the decision system is used as the task solution;

[0027] S3. Determine the maintenance process sequence based on the maintenance process information in the task solution; perform maintenance of mining equipment resources according to the maintenance process sequence, corresponding personnel, tools and spare parts, and time constraints and location constraints in the mining equipment maintenance tasks.

[0028] S1. Construct a directed graph of maintenance resources based on maintenance procedures, personnel, tools and spare parts. In the directed graph of maintenance resources, the nodes are maintenance procedures, personnel, tools and spare parts, and the edges are the dependencies between maintenance procedures, personnel, tools and spare parts. According to the resource type, convert the directed graph of maintenance resources into a hierarchical graph of the same resource type at the same level. Arrange and combine the resources at the highest level and with dependencies on each other in the hierarchical graph to obtain several root solutions. Arrange and combine the resources with dependencies on the root solutions with several root solutions in the order of the hierarchical graph from high to low to obtain several basic solutions. Delete the basic solutions with conflicts between the feasible domains of resources in several basic solutions to obtain several initial solutions.

[0029] Based on the maintenance process, and the real-time updated and accessible personnel, tools and spare parts, a directed graph of maintenance resources is constructed to clarify the relationship between resources. Then, based on the directed graph of maintenance resources, a hierarchical general graph of the same resource type at the same level is constructed, and the resources are arranged and combined in order from high to low to obtain several basic plans. Finally, the basic plans with conflicts between the feasible domains of resources are deleted, and the basic plans that can be implemented are retained to obtain several reasonable initial plans.

[0030] First, based on maintenance procedures, personnel, tools and spare parts, a maintenance resource directed graph is constructed to facilitate the clarification of the relationship between resources. Maintenance procedures, personnel, tools and spare parts are all resources. In the maintenance resource directed graph, the nodes are maintenance procedures, personnel, tools and spare parts, and the edges are the dependencies between maintenance procedures, personnel, tools and spare parts. The dependencies between maintenance procedures, personnel, tools and spare parts are specifically the dependencies between maintenance procedures and personnel, tools and spare parts, the dependencies between personnel and tools and spare parts, the dependencies between tools and spare parts, and the dependencies between maintenance procedures, personnel, tools and spare parts.

[0031] Among them, the dependencies of maintenance processes on personnel, tools and spare parts are respectively the degree of requirement of maintenance processes on personnel skills, tool necessity and spare parts necessity, the dependency between maintenance processes is the degree of requirement for maintenance sequence, the dependency between personnel is obtained by the number of times personnel appear synchronously in historical maintenance processes, the dependency between tools is obtained by the number of times tools appear synchronously in historical maintenance processes, the dependency between spare parts is obtained by the number of times tools appear synchronously in the same tool in historical maintenance processes, the dependency between personnel and tools is obtained by the number of times personnel and tools appear synchronously in historical maintenance processes, the dependency between personnel and spare parts is obtained by the number of times personnel and spare parts appear synchronously in historical maintenance processes, and the dependency between tools and spare parts is the necessity of spare parts for tools.

[0032] When assigning weights to the degree of dependency between resources, that is, the degree of dependency between personnel, tools and spare parts, that is, the degree of dependency between personnel and tools and spare parts respectively, and the degree of dependency between tools and spare parts, the assignment can be based on the following principles: the stronger the necessity for simultaneous use of resources, the greater the weight assigned to the corresponding degree of dependency; and / or the more times the resources are used simultaneously in the historical maintenance process, the greater the weight assigned to the corresponding degree of dependency; and / or experts in the maintenance field are used to assign weights to the corresponding degrees of dependency based on the degree of correlation between resource types; and / or association rules are mined between resources in historical maintenance data, and higher weights are assigned to the corresponding degrees of dependency for strongly associated resource types.

[0033] In addition, personnel, tools and spare parts have time attributes and location attributes. The time attribute indicates the time period in which the personnel, tools and spare parts are available, such as the personnel's work shifts, the idle time period of tools and spare parts, etc. The location attribute indicates the spatial location of the personnel, tools and spare parts, such as the workshop where the personnel are located, the warehouse where the tools and spare parts are located, etc. The corresponding personnel, and / or tools, and / or spare parts whose time attributes are within the time window of the task to be maintained and whose location attributes are within the location window of the task to be maintained are available. The personnel, tools and spare parts in the above-mentioned maintenance resource directed graph are all available and updated in real time.

[0034] Then, according to the resource type and the corresponding hierarchical division basis, the maintenance resource directed graph is converted into a hierarchical general graph of the same resource type at the same level. In the hierarchical general graph, several levels are arranged from high to low from top to bottom (arranged from high to low). The higher the level, the more difficult it is to easily obtain the corresponding type of resources. The resources in each level are the same type of resources, and there are still edges between resources in the hierarchical general graph, that is, there are dependencies between maintenance processes, personnel, tools, and spare parts. Converting the maintenance resource directed graph into a hierarchical general graph facilitates subsequent resource traversal based on the hierarchical relationship, and arranges and combines resources to form different solutions.

[0035] Next, arrange and combine the top-level resources in the hierarchical diagram that are dependent on each other (different substances in a mathematical sense form a combination) to obtain several root solutions; then, in the order of the hierarchical diagram from high to low, arrange and combine the resources that are dependent on the root solution with several root solutions to obtain several basic solutions. The number of resource combinations in the above arrangement and combination process can be set according to needs.

[0036] Finally, if there is a conflict between the time attribute and location attribute of the resources in the basic plan, that is, there is a conflict between the feasible domains of internal resources, in order to improve the rationality of the basic plan, the basic plan with conflicts between the feasible domains of internal resources is deleted to obtain several initial plans.

[0037] In order to further screen out suitable initial solutions, reduce the number of initial solutions, and improve computational efficiency, several initial solutions are screened to obtain several optimized initial solutions, which are used to replace several initial solutions to perform the operation in S2.

[0038] The operations of the screening process are specifically as follows: constructing an initial scoring function based on time consumption, initial time consumption weight, cost, initial cost weight, utilization rate and initial utilization rate weight; obtaining an initial scoring value of each initial solution based on the initial scoring function; taking an initial solution with an initial scoring value greater than an initial scoring value threshold as a first solution with better quality; all first solutions form a first solution set; obtaining a scoring variance of the first solution set; if the scoring variance is greater than the mean threshold, it means that there are large differences in the first solutions in the initial scoring function and the adaptability distribution is uneven, then the initial time consumption weight, and / or the initial cost weight, and / or the initial utilization rate weight are increased to the first time consumption weight, and / or the first cost weight, and / or the first utilization rate weight, respectively; based on the first time consumption weight, and / or the first cost weight, and / or the first utilization rate weight, update the initial scoring function to obtain an updated scoring function; based on the updated scoring function, obtain an updated scoring value of each initial solution, and take the initial solution with an updated scoring value greater than the updated scoring threshold as the optimized initial solution. The formula for increasing the initial time consumption weight, or the initial cost weight, or the initial utilization weight can be set according to actual parameter adjustment requirements.

[0039] In order to further improve the quality of the initial solution and further improve the calculation efficiency, if the score variance is greater than the mean threshold, and the difference between the score mean of the first solution set and the expected value of time consumption, and / or cost, and / or utilization rate is respectively less than the time consumption difference threshold, and / or the cost difference threshold, and / or the utilization rate difference threshold, at this time, the first solution is unevenly distributed with adaptability, but the initial scoring function can meet the expected requirements, then the initial time consumption weight, and / or the initial cost weight, and / or the initial utilization weight are respectively reduced to the second time consumption weight, and / or the second cost weight, and / or the second utilization weight; based on the second time consumption weight, and / or the second cost weight, and / or the second utilization weight, the initial scoring function is updated to obtain an updated scoring function; the updated scoring function is used to execute the operation of obtaining an optimized initial solution.

[0040] The operation of reducing the initial time consumption weight, or the initial cost weight, or the initial utilization weight is achieved through the following formula: w 2 = w 0 ·(1-α·(uE)) , w 2 is the second weight of time consumption, or the second weight of cost, or the second weight of utilization rate, w 0 is the initial time consumption weight, or the initial cost weight, or the initial utilization weight, α is the adjustment coefficient, u is the mean score, E It is the expected value of time consumption, or the expected value of cost, or the expected value of utilization rate.

[0041] S2. Discretize the state variables of each of the several initial plans, and arrange and combine them in the form of a decision tree to obtain a decision system containing several sub-decision plans; convert the task of mine equipment to be inspected into a task tree to be inspected in the form of a decision tree; and use the sub-decision plan in the decision system that corresponds to the maximum matching degree with the task tree to be inspected as the task solution.

[0042] After the state variables of each initial plan are discretized, they are arranged and combined in the form of a decision tree to achieve the refinement and expansion of each initial plan, and a decision system consisting of several sub-decision plans in the form of decision trees with a clear structure and easy analysis is obtained. Then, according to the matching degree between each sub-decision plan in the decision system and the task tree to be inspected, the most suitable task solution to solve the problem is screened out, thereby improving the resource utilization rate of the processing of mining equipment maintenance tasks, reducing maintenance costs, and shortening maintenance time.

[0043] First, the state variables of each of the several initial schemes are discretized. The purpose of the discretization is to convert the state variables expressed in the form of ranges in each initial scheme (such as the time attributes or location attributes of personnel, tools and spare parts expressed in the form of data ranges, for example, the time attribute of personnel in the state variable is 8:00-9:00, then the variables after discretization can be 8:00, 8:30 and 9:00; if the state variable is a specific data, the data remains unchanged before and after the discretization) into several specific data; then, the resources are further arranged and combined according to the discretized data, and different discretized data have corresponding maintenance information, which is then presented in the form of a decision tree to refine and expand the initial scheme, realize the above-mentioned arrangement and combination in the form of a decision tree, and obtain a decision system containing several sub-decision schemes.

[0044] Then, the mine equipment maintenance task is converted into a task tree in the form of a decision tree. The conversion process is the same as the way the above-mentioned initial schemes are converted into a decision system, which is also to discretize the state variables (maintenance time and maintenance location) of the mine equipment maintenance task (if the state variable of the mine equipment maintenance task is a specific data, the data remains unchanged before and after the discretization process), permutate and combine, and generate a decision tree.

[0045] Finally, if the state variable of the task to be repaired of mining equipment is a specific data, the sub-decision plan corresponding to the maximum matching degree of the task tree to be inspected in the decision system is used as the task solution. The matching degree is the product of the feature vector of the task tree to be inspected and the feature vector of the sub-decision plan. The feature vector of the task tree to be inspected is obtained based on the state variable of the task to be repaired, and the feature vector of the sub-decision plan is obtained based on the state variable of the sub-decision plan. The feature vector obtained based on the state variable can be obtained through a trained neural network, and can also be obtained based on the correspondence between the state variable and the preset vector conversion mapping table.

[0046] If the state variable of the mine equipment maintenance task is not a specific data, the sub-decision plan corresponding to the maximum value of the mean matching degree of all sub-task trees in the task tree to be inspected in the decision system is used as the task solution that best suits the mine equipment maintenance task. The matching degree is the product of the feature vector of the sub-task tree and the feature vector of the sub-decision plan. The feature vector of the sub-task tree is obtained based on the state variable of the sub-task tree, and the feature vector of the sub-decision plan is obtained based on the state variable of the sub-decision plan.

[0047] In order to cope with the crisis caused by the sudden change of the mine equipment maintenance task, if there are several task solutions and the mine equipment maintenance task changes, in order to solve the new mine equipment maintenance task, based on the mapping relationship between the state variables before the mine equipment maintenance task changes and the state variables after the change, the corresponding variables in each task solution (corresponding to the specific data obtained after discretization) are adjusted to obtain several task solution update solutions; the process execution sequence of each task solution update solution, as well as the personnel, tools and spare parts, are obtained, and the process dependency gene code and process resource gene code of each task solution update solution are obtained, and the process dependency gene code and process resource gene code of each task solution update solution are spliced ​​(the process dependency gene code and process resource gene code of the same task solution update solution are spliced) to obtain the solution scheduling code of each task solution update solution; the solution scheduling code with a fitness greater than the fitness threshold and a density less than the density threshold is used as a new task solution update solution to solve the changed mine equipment maintenance task, that is, to solve the new mine equipment maintenance task.

[0048] Among them, fitness is the sum of the coding values ​​of the solution scheduling coding; density is the mean distance between the task solution update solution and other task solution update solutions; distance is the product of the solution scheduling coding corresponding to the task solution update solution and the solution scheduling coding corresponding to other task solution update solutions, which is the current i The corresponding solution of the task solution update solution is the scheduling code of the current j The task solution update solution corresponds to the product of the scheduling code of the solution.

[0049] S3. Determine the maintenance process sequence based on the maintenance process information in the task solution; perform maintenance of mining equipment resources according to the maintenance process sequence, corresponding personnel, tools and spare parts, and time constraints and location constraints in the mining equipment maintenance tasks.

[0050] If there are several task solutions, one of them can be randomly selected to determine the order of the maintenance procedures, or the task solution corresponding to the maximum employee age can be selected to determine the order of the maintenance procedures, or the task solution corresponding to the minimum number of maintenance procedures can be selected to determine the order of the maintenance procedures.

[0051] The operation of determining the sequence of maintenance procedures can be obtained based on historical experience.

[0052] The operation of determining the maintenance process sequence can also be to establish a constraint network model, which describes the dependency constraints between the maintenance processes. By solving the constraint propagation, the maintenance process sequence that satisfies all constraints can be obtained. In order to improve the computational efficiency, the maintenance process sequence is modeled as an efficient and easy-to-execute finite state machine. The initial state S0 is the start state of the maintenance process, and the intermediate states S1...Sn correspond to each maintenance process; the state transition S0->S1->...->Sn->S (termination) describes the maintenance process sequence.

[0053] During the maintenance of mining equipment resources, hierarchical scheduling can be performed. Hierarchical scheduling includes high-level scheduling and low-level scheduling.

[0054] High-level scheduling determines the process-level order based on the maintenance process sequence, such as topological sorting.

[0055] The bottom-level scheduling assigns the same-level maintenance procedures to available resources in parallel, and assigns the corresponding maintenance procedures to the corresponding personnel according to the skill matching relevance; for tools and / or spare parts, the corresponding maintenance procedures are assigned to the corresponding personnel according to the demand degree; the bottom-level scheduling and the trip planning algorithm are used to arrange the operation route. During the execution process, if there are special circumstances, the decision can be re-triggered and the scheduling plan can be dynamically adjusted.

[0056] This embodiment also provides a system for executing a method for scheduling mining equipment maintenance resources, including:

[0057] Perceptual cells are used to construct a directed graph of maintenance resources based on maintenance procedures, personnel, tools and spare parts; in the directed graph of maintenance resources, nodes are maintenance procedures, personnel, tools and spare parts belonging to resources, and edges are dependencies between maintenance procedures, personnel, tools and spare parts; according to resource types, the directed graph of maintenance resources is converted into a hierarchical general graph of the same level and the same resource type; in the hierarchical general graph, the resources at the highest level and with dependencies on each other are arranged and combined to obtain several root solutions; in the order of the hierarchical general graph from high to low, the resources with dependencies on the root solutions are arranged and combined with several root solutions respectively to obtain several basic solutions; basic solutions with conflicts between the feasible domains of internal resources are deleted to obtain several initial solutions; perceptual cells can also be used to screen several initial solutions to obtain several optimized initial solutions. For details on the screening process, please refer to the corresponding operations in S1;

[0058] The decision cell is used to discretize the state variables of each of the initial solutions and arrange and combine them in the form of a decision tree to obtain a decision system containing several sub-decision solutions; convert the task of mine equipment to be repaired into a task tree to be inspected in the form of a decision tree; and take the sub-decision solution in the decision system corresponding to the maximum matching degree of the task tree to be inspected as the task solution;

[0059] The execution cell is used to determine the maintenance process sequence based on the maintenance process information in the task solution; according to the maintenance process sequence, the corresponding personnel, tools and spare parts, as well as the time constraints and location constraints in the mine equipment maintenance task, the mine equipment resource maintenance is carried out; the maintenance process sequence determined in the execution cell can be obtained based on historical experience, and can also be based on the establishment of a constraint network model, which describes the dependency constraints between the maintenance processes, and obtains the maintenance process sequence that satisfies all constraints through constraint propagation solution; at the same time, in order to improve the computational efficiency, the execution cell models the maintenance process sequence as an efficient and easy-to-execute finite state machine, the initial state S0 is the maintenance process start state, and the intermediate state S1...Sn corresponds to each maintenance process; the state transition S0->S1->...->Sn->S (termination) describes the maintenance process sequence; hierarchical scheduling can be performed in the process of mining equipment resource maintenance by the execution cell. Hierarchical scheduling execution includes high-level scheduling and low-level scheduling. High-level scheduling determines the process level sequence according to the maintenance process sequence, such as topological sorting. The bottom-level scheduling dispatches the same-level maintenance processes to available resources in parallel, and assigns the corresponding maintenance processes to the corresponding personnel according to the skill matching correlation; for tools and / or spare parts, the corresponding maintenance processes are assigned to the corresponding personnel according to the demand degree; the bottom-level scheduling and itinerary planning algorithm are used to arrange the work route; and during the execution process, if there are special circumstances, the decision can be re-triggered and the scheduling plan can be dynamically adjusted; through the above process, the execution cell can efficiently dispatch and execute the specific operations of the maintenance resources based on the constraint network, decision plan and state machine model to complete the maintenance task;

[0060] Evolutionary cells are used when there are several task solutions in the decision cells and the task of mine equipment to be repaired changes; based on the mapping relationship between the state variables before the change of the task of mine equipment to be repaired and the state variables after the change, the corresponding variables in each task solution are adjusted to obtain several task solution update solutions; the process execution sequence of each task solution update solution, as well as the personnel, tools and spare parts, are obtained, and the process dependency gene code and process resource gene code of each task solution update solution are obtained, and the process dependency gene code and process resource gene code of each task solution update solution are spliced ​​separately to obtain the solution scheduling code of each task solution update solution; the solution scheduling code with a fitness greater than the fitness threshold and a density less than the density threshold is used as a new task solution update solution to solve the changed mine equipment to be repaired task;

[0061] Interactive cells are used to organize and coordinate multi-cellular collaboration among perception cells, decision-making cells, and evolutionary cells based on contract networks and trust models to achieve flexible scheduling of maintenance resources. In interactive cells, multi-cellular collaboration can be carried out based on the multi-party interaction mechanism of contract networks or on the dynamic virtual team formation based on trust models.

[0062] Among them, the multi-party interaction mechanism based on the contract network is: the interactive cells use the contract network protocol to organize the division of labor and cooperation among multiple cells. Specifically, after the publishing cells used to publish tasks in the interactive cells publish the specific requirements of the tasks to be repaired by mining equipment (such as procedures, time windows, locations, etc.) on the contract network, participants such as decision-making cells and execution cells will respond to the published tasks according to their own capabilities and status, and submit plans to share the task requirements; the publisher selects the best plan from the participants' responses, signs a contract with the corresponding cells, and determines the obligations and rights of both parties. At the same time, during the execution of the contract, the supervisory cells in the interactive cells are responsible for supervising whether the decision-making cells, execution cells, etc. fulfill their obligations in accordance with the terms of the contract, and handle them in case of breach of contract.

[0063] Through the mutual collaboration among publishing cells, participants (decision-making cells, execution cells, etc.) and supervisory cells, interactive cells can flexibly organize multiple cells to divide the work and collaborate on demand to dynamically meet maintenance needs.

[0064] Dynamic virtual team formation based on trust model: Interactive cells also introduce a trust model to quantify and manage the trust relationship between multiple cells. Based on the historical cooperation records between cells, the trust scores between cells are calculated. According to the current requirements of the mine equipment maintenance tasks, cells (decision cells, execution cells, etc.) with mutual trust greater than the trust threshold are selected from the cell group to form a dynamic virtual collaboration team. Interactive cells will self-learn and adjust the trust according to the actual collaboration effect of the virtual collaboration team, and continuously optimize the trust model. Through the trust model, interactive cells can flexibly dispatch high-trust cells to form an efficient collaboration team, thereby improving the coordination and reliability of maintenance resource scheduling.

[0065] In general, interactive cells, based on contract networks and trust models, achieve orderly interaction and dynamic collaborative organization among multiple cells, avoid the drawbacks of single-center scheduling, and improve the flexibility and robustness of resource scheduling.

[0066] In order to verify the effectiveness of the system corresponding to the method of this embodiment, the following experiment was conducted.

[0067] There are 70 personnel nodes, 50 tool nodes, and 200 spare parts nodes in the maintenance resource directed graph of the perception cell. The initial scoring function is set to F = 0.5 cost + 0.3 time + 0.2 utilization. The decision cell extracts 25 key state variables, the average value domain size is 5, and about 10 6 The decision-making system of the leaf node. The execution cell is modeled as a finite state machine with 122 state nodes, corresponding to 122 maintenance processes; the high-level scheduling uses a topological sorting algorithm to determine the order of the processes, and the low-level scheduling uses a genetic algorithm to solve the traveling salesman problem and arrange specific work routes. The contract network in the interactive cell includes 1 publisher cell (enterprise decision-making layer), 10-15 participants (field team), and 3 supervisory cells (quality inspectors); the initial trust score of the trust model is 0.5 for all members, and the new members inherit the average score of the team, and the score of the old members = 0.9 old score + 0.1 new score. There are 200 task solution update solutions in the evolutionary cell, with a coding length of 122, corresponding to 122 processes.

[0068] The test data includes 100 mine instances, 500,000 processes, and 2 million resources (maintenance processes, personnel, tools, and spare parts); the test environment is a high-performance workstation, and the algorithm is implemented in parallel in C++. The test results are as follows: the maintenance cost of the traditional manual scheduling result is 1.24 million yuan, the maintenance time is 26 days, and the resource utilization rate is 65%; the maintenance cost of the genetic algorithm is 1.07 million yuan, the maintenance time is 21 days, and the resource utilization rate is 76%; the maintenance cost of the ant algorithm is 1.1 million yuan, the maintenance time is 22 days, and the resource utilization rate is 73%; the maintenance cost of this embodiment is 1.03 million yuan, the maintenance time is 19 days, and the resource utilization rate is 81%; compared with the manual method, this embodiment can reduce the maintenance cost by 17%, shorten the maintenance time by 27%, and improve the resource utilization rate by 24% (significant difference value p<0.01, significant difference). In addition, in the experimental results of the convergence and time complexity of this embodiment, the typical convergence time of this embodiment is 4 hours, and the longest convergence time for a large-scale 1000 variable instance is 10 hours, which meets the real-time requirements of the project, has controllable time complexity and good stability.

[0069] In summary, the system corresponding to the method of this embodiment can efficiently and reliably solve the problem of complex mine maintenance resource scheduling, and has practical value. It proves that the method of this embodiment can reduce costs and maintenance time and improve resource utilization in mine equipment maintenance resource scheduling.

[0070] This embodiment further provides a scheduling device for mining equipment maintenance resources, including a processor and a memory, wherein the processor implements the above-mentioned scheduling method for mining equipment maintenance resources when executing a computer program stored in the memory.

[0071] This embodiment further provides a computer-readable storage medium for storing a computer program, wherein the computer program implements the above-mentioned method for scheduling mining equipment maintenance resources when executed by a processor.

[0072] The present embodiment provides a method for scheduling mining equipment maintenance resources. First, based on the maintenance process, and personnel, tools and spare parts that are updated in real time and available, a directed graph of maintenance resources is constructed to facilitate clarifying the relationship between resources. Then, based on the directed graph of maintenance resources, a hierarchical general graph of the same resource type at the same level is constructed, and the resources are arranged and combined in order from high to low to obtain several basic solutions, and the basic solutions with conflicts between the feasible domains of resources are deleted to obtain several reasonable initial solutions. Then, the state variables of each initial solution are discretized and arranged and combined in the form of a decision tree to achieve refinement and expansion of each initial solution, and a decision system consisting of several sub-decision solutions in the form of decision trees with a clear structure and easy to analyze is obtained, and the task solution most suitable for solving the problem is screened out according to the matching degree between each sub-decision solution in the decision system and the task tree to be inspected; finally, the maintenance process sequence is determined based on the task solution, and the mining equipment resource maintenance is performed; the method can reduce the maintenance cost, shorten the maintenance time, and improve the resource utilization rate, and has good calculation stability.

Claims

1. A method for scheduling mining equipment maintenance resources. It is characterized in that The following operations are included: S1. Based on maintenance procedures, personnel, tools and spare parts, a maintenance resource directed graph is constructed; maintenance resources include maintenance procedures, personnel, tools and spare parts; in the maintenance resource directed graph, nodes are maintenance procedures, personnel, tools and spare parts, and edges are dependencies between maintenance procedures, personnel, tools and spare parts; Dependencies between maintenance processes, personnel, tools, and spare parts, specifically, the dependencies between maintenance processes and personnel, tools, and spare parts, the dependencies between personnel and tools and spare parts, the dependencies between tools and spare parts, and the dependencies between maintenance processes, personnel, tools, and spare parts; The dependency relationships between maintenance processes and personnel, tools and spare parts are respectively the degree of requirement of maintenance processes on personnel skills, necessity of tools and necessity of spare parts; the dependency relationship between maintenance processes is the degree of requirement on maintenance sequence; the dependency relationship between personnel is obtained by the number of times personnel appear simultaneously in the historical maintenance process; The dependency between tools is obtained from the number of times the tools appear synchronously in the historical maintenance process; the dependency between spare parts is obtained from the number of times the tools appear synchronously in the same tool in the historical maintenance process; The dependency between personnel and tools is obtained by the number of times personnel and tools appear simultaneously in the historical maintenance process; the dependency between personnel and spare parts is obtained by the number of times personnel and spare parts appear simultaneously in the historical maintenance process; the dependency between tools and spare parts is the degree of necessity of spare parts for tools; According to the resource type, the maintenance resource directed graph is converted into a hierarchical general graph of the same resource type at the same level; Arrange and combine the resources of the highest level and having dependencies with each other in the hierarchical general graph to obtain several root solutions; arrange and combine the resources having dependencies with the root solutions with several root solutions in the order of the hierarchical general graph from high to low to obtain several basic solutions; delete the basic solutions with conflicts between the feasible domains of internal resources to obtain several initial solutions; S2. After discretizing the state variables of each of the initial solutions, the state variables are arranged and combined in the form of a decision tree to obtain a decision system containing several sub-decision solutions; the state variables of the initial solutions are the time attributes and location attributes of personnel, tools and spare parts; Convert the tasks of mine equipment to be inspected and repaired into a task tree in the form of a decision tree; The conversion process is: discretizing and arranging the state variables of the tasks to be repaired of the mining equipment, and generating a decision tree; the state variables of the tasks to be repaired are the repair time and the repair location; If the state variable of the task to be repaired of the mining equipment is a specific data, the sub-decision plan corresponding to the maximum matching degree of the task tree to be repaired in the decision system is used as the task solution; the matching degree is the product of the feature vector of the task tree to be repaired and the feature vector of the sub-decision plan; the feature vector of the task tree to be repaired is obtained based on the state variable of the task to be repaired, and the feature vector of the sub-decision plan is obtained based on the state variable of the sub-decision plan; If the state variable of the task to be inspected and maintained of the mining equipment is not a specific data, the sub-decision plan corresponding to the maximum value of the mean value of the matching degree of all sub-task trees in the task tree to be inspected in the decision system is used as the task solution; the matching degree is the product of the characteristic vector of the sub-task tree and the characteristic vector of the sub-decision plan, the characteristic vector of the sub-task tree is obtained based on the state variable of the sub-task tree, and the characteristic vector of the sub-decision plan is obtained based on the state variable of the sub-decision plan; S3. Determine the maintenance process sequence based on the maintenance process information in the task solution; perform maintenance of mining equipment resources according to the maintenance process sequence, corresponding personnel, tools and spare parts, and time constraints and location constraints in the mining equipment maintenance task.

2. The method for scheduling mining equipment maintenance resources according to claim 1, It is characterized in that Screening several initial solutions to obtain several optimized initial solutions, which are used to replace several initial solutions to perform the operation in S2; The operations of the screening process are specifically as follows: Based on the time consumption, initial time consumption weight, cost, initial cost weight, utilization rate and initial utilization rate weight, an initial scoring function is constructed; based on the initial scoring function, an initial scoring value of each initial solution is obtained; An initial solution whose initial score value is greater than the initial score value threshold is taken as the first solution; all the first solutions form a first solution set; and the score variance of the first solution set is obtained; If the score variance is greater than the mean threshold, the initial time consumption weight, and / or the initial cost weight, and / or the initial utilization weight are increased to the first time consumption weight, and / or the first cost weight, and / or the first utilization weight, respectively; Based on the first weight of time consumption, and / or the first weight of cost, and / or the first weight of utilization rate, updating the initial scoring function to obtain an updated scoring function; Based on the update scoring function, the update scoring value of each initial solution is obtained, and the initial solution with an update scoring value greater than the update scoring threshold is used as the optimized initial solution.

3. The method for scheduling mining equipment maintenance resources according to claim 1, It is characterized in that If there are several task solutions in S2, and the task of mine equipment to be repaired changes; Based on the mapping relationship between the state variables of the mine equipment maintenance task before and after the change, the corresponding variables in each task solution are adjusted to obtain several task solution update solutions; Obtain the process execution sequence of each task solution update plan, as well as personnel, tools and spare parts, perform coding processing, obtain the process dependency gene code and process resource gene code of each task solution update plan, and splice the process dependency gene code and process resource gene code of each task solution update plan respectively to obtain the solution scheduling code of each task solution update plan; The scheme scheduling code with fitness greater than the fitness threshold and density less than the density threshold is used as the new task solution update solution to solve the changed mining equipment maintenance tasks; The fitness is the sum of the coding values ​​of the solution scheduling coding; the density is the mean distance between the task solution update solution and other task solution update solutions; and the distance is the product of the solution scheduling coding and other solution scheduling coding.

4. A system for executing the method for scheduling mining equipment maintenance resources according to claim 1, It is characterized in that include: The sensor cells are used to construct a directed graph of maintenance resources based on maintenance procedures, personnel, tools and spare parts; maintenance resources include maintenance procedures, personnel, tools and spare parts; in the directed graph of maintenance resources, the nodes are maintenance procedures, personnel, tools and spare parts, and the edges are the dependencies between maintenance procedures, personnel, tools and spare parts; the dependencies between maintenance procedures, personnel, tools and spare parts are specifically the dependencies between maintenance procedures and personnel, tools and spare parts, the dependencies between personnel and tools and spare parts, the dependencies between tools and spare parts, and the dependencies between maintenance procedures, personnel, tools and spare parts; the dependencies between maintenance procedures and personnel, tools and spare parts are respectively the requirements of maintenance procedures for personnel skills, tool necessity and spare parts necessity; the dependencies between maintenance procedures are the requirements for maintenance sequence; the dependencies between personnel are obtained by the number of times personnel appear simultaneously in the historical maintenance process; The dependency between tools is obtained from the number of times the tools appear synchronously in the historical maintenance process; the dependency between spare parts is obtained from the number of times the tools appear synchronously in the same tool in the historical maintenance process; The dependency between personnel and tools is obtained by the number of times personnel and tools appear synchronously in the historical maintenance process; the dependency between personnel and spare parts is obtained by the number of times personnel and spare parts appear synchronously in the historical maintenance process; the dependency between tools and spare parts is the degree of necessity of spare parts for tools; according to the resource type, the maintenance resource directed graph is converted into a hierarchical general graph of the same resource type at the same level; Arrange and combine the resources of the highest level and having dependencies with each other in the hierarchical general graph to obtain several root solutions; arrange and combine the resources having dependencies with the root solutions with several root solutions in the order of the hierarchical general graph from high to low to obtain several basic solutions; delete the basic solutions with conflicts between the feasible domains of internal resources to obtain several initial solutions; Decision cells are used to discretize the state variables of each of several initial solutions and arrange and combine them in the form of a decision tree to obtain a decision system containing several sub-decision solutions; the state variables of the initial solution are the time attributes and location attributes of personnel, tools and spare parts; the tasks to be repaired of mining equipment are converted into a task tree to be repaired in the form of a decision tree; The conversion process is as follows: discretizing and arranging the state variables of the task to be repaired of the mining equipment, and generating a decision tree; the state variables of the task to be repaired are the repair time and the repair location; if the state variable of the task to be repaired of the mining equipment is a specific data, then the sub-decision scheme corresponding to the maximum matching degree of the task tree to be inspected in the decision system is used as the task solution; the matching degree is the product of the characteristic vector of the task tree to be inspected and the characteristic vector of the sub-decision scheme; the characteristic vector of the task tree to be inspected is obtained based on the state variable of the task to be repaired, and the characteristic vector of the sub-decision scheme is obtained based on the state variable of the sub-decision scheme; if the state variable of the task to be repaired of the mining equipment is not a specific data, then the sub-decision scheme corresponding to the maximum matching degree mean value of all sub-task trees in the task tree to be inspected in the decision system is used as the task solution; the matching degree is the product of the characteristic vector of the sub-task tree and the characteristic vector of the sub-decision scheme, the characteristic vector of the sub-task tree is obtained based on the state variable of the sub-task tree, and the characteristic vector of the sub-decision scheme is obtained based on the state variable of the sub-decision scheme; The execution cell is used to determine the maintenance process sequence based on the maintenance process information in the task solution; according to the maintenance process sequence, the corresponding personnel, tools and spare parts, as well as the time constraints and location constraints in the mine equipment maintenance tasks, the mine equipment resources are maintained.

5. A dispatching device for mining equipment maintenance resources, It is characterized in that It comprises a processor and a memory, wherein when the processor executes the computer program stored in the memory, the method for scheduling mining equipment maintenance resources as described in any one of claims 1 to 3 is implemented.

6. A computer-readable storage medium, It is characterized in that Used to store a computer program, wherein when the computer program is executed by a processor, the method for scheduling mining equipment maintenance resources according to any one of claims 1 to 3 is implemented.