Maintenance resource optimization method considering maintenance task

Abstract

The invention belongs to the field of weaponry maintenance support technology optimization, and particularly relates to a maintenance resource optimization method considering maintenance tasks. The method comprises the steps of 1, determining a calculation model of the combat effectiveness E of a weapon equipment system; 2, determining an objective function calculation model; and step 3, optimizing the objective function based on a genetic algorithm. Through the implementation of the method, a weapon equipment system combat effectiveness calculation model is constructed, an optimization objective function is determined, and an optimal weapon equipment maintenance support resource configuration result is obtained through iterative solution by adopting a genetic algorithm.

Description

Technical Field

[0001] The invention belongs to the field of weapon equipment maintenance support technology optimization, and specifically relates to a maintenance resource optimization method taking maintenance tasks into consideration. Background Art

[0002] Maintenance support resources are the material basis necessary for the weapon equipment support system to implement maintenance support tasks, and are the basic carrier for carrying out maintenance support activities. The shortage of types or quantities of maintenance support resources directly affects the maintenance support capabilities of the maintenance support organization, thereby affecting the smooth implementation of the maintenance tasks of the corresponding maintenance support organization. When multiple maintenance support tasks need to be performed at the same time, how to reasonably allocate limited maintenance support resources to each maintenance support task is a common problem in maintenance support research. Therefore, the reasonable allocation of maintenance support resources has become the core and difficulty of the digital operation and maintenance and support service system, and has also become one of the key directions of research on the combat application of weapon equipment systems. At present, particle swarm algorithms and genetic algorithms are commonly used in the research on equipment maintenance support resource optimization, with the goal of minimizing the average maintenance time and minimizing the maintenance support cost, without considering the effectiveness of maintenance support. Summary of the invention

[0003] 1. Technical issues to be resolved

[0004] The technical problem to be solved by the present invention is: how to propose a maintenance resource optimization method taking maintenance tasks into consideration, so as to maximize the combat effectiveness of the weapon equipment system under the premise of completing the maintenance support tasks.

[0005] (II) Technical solution

[0006] In order to solve the above technical problems, the present invention provides a maintenance resource optimization method considering maintenance tasks, and the maintenance resource optimization method considering maintenance tasks includes:

[0007] Step 1: Determine the calculation model of the combat effectiveness E of the weapon and equipment system;

[0008] Step 2: Determine the objective function calculation model;

[0009] Step 3: Optimize the objective function based on genetic algorithm;

[0010] Among them, in step 1, the calculation model of the combat effectiveness E of the weapon and equipment system is determined;

[0011] The combat effectiveness E of the weapon and equipment system is related to the weight of each maintenance and support task and the combat effectiveness of the weapon and equipment under the resource allocation plan of the maintenance and support task;

[0012]

[0013] In step 2, the objective function calculation model is determined;

[0014] The purpose of the weapon equipment maintenance support resource allocation optimization method is to maximize the combat effectiveness E of the weapon equipment system while ensuring the completion of the maintenance support task; the optimization objective function is expressed as:

[0015] f=Max E

[0016]

[0017] Among them, U j1 is the number of the first type of maintenance resources in the resource allocation plan for the jth maintenance support task, U jm is the number of the mth type of maintenance resources in the resource allocation plan for the jth maintenance support task;

[0018] In step 3, the objective function is optimized based on the genetic algorithm.

[0019] Genetic algorithm is used to optimize the objective function to solve the problem of allocating limited maintenance support resources to multiple maintenance support tasks.

[0020] The process of optimizing the objective function based on the genetic algorithm in step 3 includes:

[0021] Step 301: Initialize genetic algorithm parameters;

[0022] Step 302: Select decision variables of the genetic algorithm;

[0023] Step 303: Select the value range and constraint conditions of the decision variables;

[0024] Step 304: Generate an initial population;

[0025] Step 305: Calculate the objective function value and the constraint function value;

[0026] Step 306: assigning fitness function values;

[0027] Step 307: Punishment;

[0028] Step 308: Determine whether the termination condition is met;

[0029] Step 309: selection;

[0030] Step 310: cross;

[0031] Step 311: mutation;

[0032] Step 312: Re-insert;

[0033] After executing step 312, go to step 305 for execution;

[0034] Steps 305 to 312 are repeated until the termination condition is met.

[0035] In step 301, the genetic algorithm parameters are initialized;

[0036] Initialization genetic algorithm parameters include: population size P s , the maximum genetic generation G max , the number of maintenance support tasks N, the generation gap GGAP, the mutation probability P m and the crossover probability P cr ;

[0037] Population size P s , mutation probability P m With the crossover probability P cr is the control parameter that affects the optimization effect in the genetic algorithm, P m The larger the value, the worse the result, because P m The larger the value, the greater the probability of generating new individuals, and the more similar it is to the random search algorithm, so P m The value is very small, 0.01; because P cr The smaller the value, the greater the probability of survival of individuals with high fitness, which is equivalent to a reduction in the solution space and a worse result. By comparing the optimization results under different numbers of maintenance and support tasks, it is concluded that when the number of maintenance and support tasks changes, P needs to be set according to the number of maintenance and support tasks. s , both can be P m =0.01, P cr ≥0.7 can ensure that the optimization results will not deteriorate.

[0038] In step 302, a decision variable of a genetic algorithm is selected;

[0039] Select the sequence numbers of the resource allocation schemes for N maintenance support tasks as decision variables. The N decision variables form an N-dimensional row resource allocation scheme vector:

[0040] spliting=[number1,number2,...,number N ].

[0041] In step 303, the value range and constraint conditions of the decision variables are selected;

[0042] For the basis vector BaseV of the decision variable, an M-ary basis vector is used, 1 indicates that the first executable resource allocation scheme is selected for the j-th maintenance support task, and M indicates that the M-th executable resource allocation scheme is selected for the j-th maintenance support task.

[0043] In the step 304, an initial population is generated;

[0044] Randomly generate P s The initial population of row resource allocation schemes with a length of N and a basis vector of BaseV; thus, the initial population formed is P s * N matrix.

[0045] In step 305, the objective function value and the constraint condition function value are calculated;

[0046] Calculate P s The combat effectiveness E and constraint function value corresponding to the N-dimensional resource configuration plan vector.

[0047] In the step 306, a fitness function value is assigned;

[0048] A fitness allocation algorithm based on linear sorting is used to assign fitness values ​​to the population. The pressure difference is selected as 2. The smaller the objective function value, the larger the fitness value.

[0049] In step 307, punishment is performed;

[0050] Penalize the fitness function value of the resource allocation solution vector that does not meet the constraints in the population to reduce its fitness value; use a smaller penalty factor in the early stage of iteration to allow more solutions to participate in the evolution; use a larger penalty factor in the later stage of iteration to ensure the quality of the solution;

[0051] In step 308, it is determined whether the termination condition is met. If so, the optimal resource allocation solution vector and the optimized combat effectiveness E are output; if not, step 309 is executed; the termination condition is: the maximum genetic generation number G is reached max Or when the optimal resource allocation solution vector does not change during the evolution process;

[0052] In the step 309, a selection is made;

[0053] The two groups of populations are selected based on the fitness function value; the random traversal sampling operator is used, and the generation gap GGAP=0.9;

[0054] In the step 310, cross;

[0055] Crossover is to generate a new row resource allocation scheme vector by exchanging some parts of two row resource allocation scheme vectors in the resource allocation scheme vector population; advanced recombination operators and single-point crossover operators are used for crossover processing;

[0056] In the step 311, mutation;

[0057] The mutation is to generate a new row resource configuration scheme vector by replacing a part of a row resource configuration scheme vector in the resource configuration scheme vector population with other parts of the row resource configuration scheme vector; a discrete mutation operator is used to perform the mutation operation;

[0058] In the step 312, reinsert;

[0059] Fitness-based reinsertion is used to ensure that P s ×(1-GGAP) most adapted split vectors are always propagated continuously to the next generation;

[0060] After executing step 312, go to step 305 for execution;

[0061] Steps 305 to 312 are repeated until a termination condition is met.

[0062] (III) Beneficial effects

[0063] Compared with the prior art, the present invention aims at the problem of reasonable allocation of limited weapon equipment system support resources and multiple maintenance support tasks, and proposes a maintenance resource optimization method taking maintenance tasks into consideration, so as to maximize the combat effectiveness of the weapon equipment system under the premise of completing the maintenance support tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a flow chart of the technical solution of the present invention.

[0065] Figure 2 This is a flow chart of a maintenance resource optimization method considering maintenance tasks according to an embodiment of the present invention. DETAILED DESCRIPTION

[0066] In order to make the purpose, content, and advantages of the present invention more clear, the specific implementation methods of the present invention are further described in detail below in conjunction with the accompanying drawings and examples.

[0067] In order to solve the above technical problems, the present invention provides a maintenance resource optimization method considering maintenance tasks, wherein the maintenance resource optimization method considering maintenance tasks is used to maximize the combat effectiveness of a weapon equipment system under the premise of completing maintenance support tasks;

[0068] The implementation process of the weapon equipment maintenance support resource allocation optimization method is as follows:

[0069] Let the maintenance support task set be T = {T j}={T1,T2,...,T N}, N is the number of maintenance support tasks, T j is the jth maintenance support task;

[0070] Maintenance support resources are U = {Ui *n i}={U1*n1,U2*n2,...U m *n m}, where: U i is the i-th type of maintenance resource, n i is the number of maintenance resources of the i-th type, and m is the number of maintenance resource types;

[0071] The maintenance resource allocation scheme set is Y = {Y j}={Y1,Y2,..,Y N}, the maintenance resource allocation scheme set for each j maintenance support task is Y j ={y jk}={y j1 ,y j2 ,..,y jM},y jk is the kth executable configuration scheme among the jth maintenance support task resource configuration schemes, M is the total number of executable resource configuration schemes, and the same resource may be used in multiple resource configuration schemes;

[0072] The combat effectiveness of weapons and equipment under each configuration scheme is different. jk The corresponding performance value is The weight of the jth maintenance support task is w j , the weights of each maintenance support task satisfy The combat effectiveness of the weapon and equipment system is E;

[0073] Therefore, the problem to be solved by the maintenance resource optimization method considering maintenance tasks is transformed into y jk How to choose the weapon and equipment system that can maximize its combat effectiveness E while ensuring the completion of maintenance and support tasks.

[0074] Table 1 is a schematic diagram of a maintenance support task and a set of applicable resource configuration solutions according to an embodiment of the present invention.

[0075] Wherein, the maintenance resource optimization method considering maintenance tasks includes:

[0076] Step 1: Determine the calculation model of the combat effectiveness E of the weapon and equipment system;

[0077] Step 2: Determine the objective function calculation model;

[0078] Step 3: Optimize the objective function based on genetic algorithm;

[0079] Among them, in step 1, the calculation model of the combat effectiveness E of the weapon and equipment system is determined;

[0080] The combat effectiveness E of the weapon and equipment system is related to the weight of each maintenance and support task and the combat effectiveness of the weapon and equipment under the resource allocation plan of the maintenance and support task;

[0081]

[0082] In step 2, the objective function calculation model is determined;

[0083] The purpose of the weapon equipment maintenance support resource allocation optimization method is to maximize the combat effectiveness E of the weapon equipment system while ensuring the completion of the maintenance support task; the optimization objective function is expressed as:

[0084] f=Max E

[0085]

[0086] Among them, U j1 is the number of the first type of maintenance resources in the resource allocation plan for the jth maintenance support task, U jm is the number of the mth type of maintenance resources in the resource allocation plan for the jth maintenance support task;

[0087] In step 3, the objective function is optimized based on the genetic algorithm.

[0088] Genetic algorithm (GA) is a bionic algorithm that overcomes the shortcoming of traditional search methods that fall into local optimal solutions and has good global search performance. Its unique operating mechanism enables it to solve large and complex combinatorial optimization problems with a limited search space at a limited cost; therefore, genetic algorithms are widely used in engineering optimization; the present invention uses genetic algorithms to optimize the objective function and solve the problem of allocating limited maintenance and support resources to multiple maintenance and support tasks.

[0089] Figure 2 is a flow chart of a maintenance resource optimization method considering maintenance tasks according to an embodiment of the present invention;

[0090] The process of optimizing the objective function based on the genetic algorithm in step 3 includes:

[0091] Step 301: Initialize genetic algorithm parameters;

[0092] Step 302: Select decision variables of the genetic algorithm;

[0093] Step 303: Select the value range and constraint conditions of the decision variables;

[0094] Step 304: Generate an initial population;

[0095] Step 305: Calculate the objective function value and the constraint function value;

[0096] Step 306: assigning fitness function values;

[0097] Step 307: Punishment;

[0098] Step 308: Determine whether the termination condition is met;

[0099] Step 309: selection;

[0100] Step 310: cross;

[0101] Step 311: mutation;

[0102] Step 312: Re-insert;

[0103] After executing step 312, go to step 305 for execution;

[0104] Steps 305 to 312 are repeated until a termination condition is met.

[0105] In step 301, the genetic algorithm parameters are initialized;

[0106] Initialization genetic algorithm parameters include: population size P s , the maximum genetic generation G max , the number of maintenance support tasks N, the generation gap GGAP, the mutation probability P m and the crossover probability P cr ;

[0107] Population size P s , mutation probability P m With the crossover probability P cr is the control parameter that affects the optimization effect in the genetic algorithm, P m The larger the value, the worse the result, because P m The larger the value, the greater the probability of generating new individuals, and the more similar it is to the random search algorithm, so P m The value is very small, 0.01 can be taken; because P cr The smaller the value, the greater the probability of survival of individuals with high fitness, which is equivalent to a reduction in the solution space and a worse result. By comparing the optimization results under different numbers of maintenance and support tasks, it is concluded that when the number of maintenance and support tasks changes, P needs to be set according to the number of maintenance and support tasks. s , both can be P m =0.01, P cr ≥0.7 can ensure that the optimization results will not deteriorate.

[0108] In this embodiment, the generation gap GGAP = 0.9, P m =0.01, P cr =0.7.

[0109] In step 302, a decision variable of a genetic algorithm is selected;

[0110] Select the sequence numbers of the resource allocation schemes for N maintenance support tasks as decision variables. The N decision variables form an N-dimensional row resource allocation scheme vector:

[0111] spliting=[number1,number2,...,number N ].

[0112] In step 303, the value range and constraint conditions of the decision variables are selected;

[0113] For the basis vector BaseV of the decision variable, an M-ary basis vector is used, 1 indicates that the first executable resource allocation scheme is selected for the j-th maintenance support task, and M indicates that the M-th executable resource allocation scheme is selected for the j-th maintenance support task.

[0114] In the step 304, an initial population is generated;

[0115] Randomly generate P s The initial population of row resource allocation schemes with a length of N and a basis vector of BaseV; thus, the initial population formed is P s * N matrix.

[0116] In step 305, the objective function value and the constraint condition function value are calculated;

[0117] Calculate P s The combat effectiveness E and constraint function value corresponding to the N-dimensional resource configuration plan vector.

[0118] In the step 306, a fitness function value is assigned;

[0119] A fitness allocation algorithm based on linear sorting is used to assign fitness values ​​to the population. The pressure difference is selected as 2. The smaller the objective function value, the larger the fitness value.

[0120] In step 307, punishment is performed;

[0121] Penalize the fitness function value of the resource allocation solution vector that does not meet the constraints in the population to reduce its fitness value; use a smaller penalty factor in the early stage of iteration to allow more solutions to participate in the evolution; use a larger penalty factor in the later stage of iteration to ensure the quality of the solution;

[0122] In step 308, it is determined whether the termination condition is met. If so, the optimal resource allocation solution vector and the optimized combat effectiveness E are output; if not, step 309 is executed; the termination condition is: the maximum genetic generation number G is reached maxOr when the optimal resource allocation solution vector does not change during the evolution process;

[0123] In the step 309, a selection is made;

[0124] The two groups of populations are selected based on the fitness function value; the random traversal sampling operator is used, and the generation gap GGAP=0.9;

[0125] In the step 310, cross;

[0126] Crossover is to generate a new row resource allocation scheme vector by exchanging some parts of two row resource allocation scheme vectors in the resource allocation scheme vector population; advanced recombination operators and single-point crossover operators are used for crossover processing;

[0127] In the step 311, mutation;

[0128] The mutation is to generate a new row resource configuration scheme vector by replacing a part of a row resource configuration scheme vector in the resource configuration scheme vector population with other parts of the row resource configuration scheme vector; a discrete mutation operator is used to perform the mutation operation;

[0129] In the step 312, reinsert;

[0130] Fitness-based reinsertion is used to ensure that P s ×(1-GGAP) most adapted split vectors are always propagated continuously to the next generation;

[0131] After executing step 312, go to step 305 for execution;

[0132] Steps 305 to 312 are repeated until the termination condition is met.

[0133] Table 1 Examples of maintenance support tasks and applicable maintenance support resource allocation schemes

[0134]

[0135]

[0136] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A maintenance resource optimization method considering maintenance tasks, characterized in that: The weapon equipment maintenance support resource allocation optimization method comprises: Step 1: Determine the calculation model of the combat effectiveness E of the weapon and equipment system; Step 2: Determine the objective function calculation model; Step 3: Optimize the objective function based on genetic algorithm.

2. The maintenance resource optimization method considering maintenance tasks as claimed in claim 1, characterized in that: In the step 1, a calculation model of the combat effectiveness E of the weapon and equipment system is determined; The combat effectiveness E of the weapon and equipment system is related to the weight of each maintenance and support task and the combat effectiveness of the weapon and equipment under the resource allocation plan of the maintenance and support task; In the step 2, determining the objective function calculation model; The purpose of the weapon equipment maintenance support resource allocation optimization method is to maximize the combat effectiveness E of the weapon equipment system while ensuring the completion of the maintenance support task; the optimization objective function is expressed as: f=Max E Among them, U j1 is the number of the first type of maintenance resources in the resource allocation plan for the jth maintenance support task, U jm is the number of the mth type of maintenance resources in the resource allocation plan for the jth maintenance support task; In the step 3, the objective function is optimized based on a genetic algorithm; Genetic algorithm is used to optimize the objective function to solve the problem of allocating limited maintenance support resources to multiple maintenance support tasks.

3. The maintenance resource optimization method considering maintenance tasks as claimed in claim 2, characterized in that: The process of optimizing the objective function based on the genetic algorithm in step 3 includes: Step 301: Initialize genetic algorithm parameters; Step 302: Select decision variables of the genetic algorithm; Step 303: Select the value range and constraint conditions of the decision variables; Step 304: Generate an initial population; Step 305: Calculate the objective function value and the constraint function value; Step 306: assigning fitness function values; Step 307: Punishment; Step 308: Determine whether the termination condition is met; Step 309: selection; Step 310: cross; Step 311: mutation; Step 312: Re-insert; After executing step 312, go to step 305 for execution; Steps 305 to 312 are repeated until a termination condition is met.

4. The maintenance resource optimization method considering maintenance tasks as claimed in claim 3, characterized in that: In step 301, the genetic algorithm parameters are initialized; Initialization genetic algorithm parameters include: population size P s , the maximum genetic generation G max , the number of maintenance support tasks N, the generation gap GGAP, the mutation probability P m and the crossover probability P cr ; Population size P s , mutation probability P m With the crossover probability P cr is the control parameter that affects the optimization effect in the genetic algorithm, P m The larger the value, the worse the result, because P m The larger the value, the greater the probability of generating new individuals, and the more similar it is to the random search algorithm, so P m The value is very small, 0.01; because P cr The smaller the value, the greater the probability of survival of individuals with high fitness, which is equivalent to a reduction in the solution space and a worse result. By comparing the optimization results under different numbers of maintenance and support tasks, it is concluded that when the number of maintenance and support tasks changes, P needs to be set according to the number of maintenance and support tasks. s , both can be P m =0.01, P cr ≥0.7 can ensure that the optimization results will not deteriorate.

5. The maintenance resource optimization method considering maintenance tasks as claimed in claim 4, characterized in that: In step 302, a decision variable of a genetic algorithm is selected; Select the sequence numbers of the resource allocation schemes for N maintenance support tasks as decision variables. The N decision variables form an N-dimensional row resource allocation scheme vector: spliting=[number1,number2,...,number N ]。 6. The maintenance resource optimization method considering maintenance tasks as claimed in claim 5, characterized in that: In step 303, the value range and constraint conditions of the decision variables are selected; For the basis vector BaseV of the decision variable, an M-ary basis vector is used, 1 indicates that the first executable resource allocation scheme is selected for the j-th maintenance support task, and M indicates that the M-th executable resource allocation scheme is selected for the j-th maintenance support task.

7. The maintenance resource optimization method considering maintenance tasks as claimed in claim 6, characterized in that: In the step 304, an initial population is generated; Randomly generate P s The initial population of row resource allocation schemes with a length of N and a basis vector of BaseV; thus, the initial population formed is P s * N matrix.

8. The maintenance resource optimization method considering maintenance tasks as claimed in claim 7, characterized in that: In step 305, the objective function value and the constraint condition function value are calculated; Calculate P s The combat effectiveness E and constraint function value corresponding to the N-dimensional resource configuration plan vector.

9. The maintenance resource optimization method considering maintenance tasks as claimed in claim 8, characterized in that: In the step 306, a fitness function value is assigned; A fitness allocation algorithm based on linear sorting is used to assign fitness values ​​to the population. The pressure difference is selected as 2. The smaller the objective function value, the larger the fitness value.

10. The maintenance resource optimization method considering maintenance tasks according to claim 9, characterized in that: In step 307, punishment is performed; Penalize the fitness function value of the resource allocation solution vector that does not meet the constraints in the population to reduce its fitness value; use a smaller penalty factor in the early stage of iteration to allow more solutions to participate in the evolution; use a larger penalty factor in the later stage of iteration to ensure the quality of the solution; In step 308, it is determined whether the termination condition is met. If so, the optimal resource allocation solution vector and the optimized combat effectiveness E are output; if not, step 309 is executed; the termination condition is: the maximum genetic generation number G is reached max Or when the optimal resource allocation solution vector does not change during the evolution process; In the step 309, a selection is made; The two groups of populations are selected based on the fitness function value; the random traversal sampling operator is used, and the generation gap GGAP=0.9; In the step 310, cross; Crossover is to generate a new row resource allocation scheme vector by exchanging some parts of two row resource allocation scheme vectors in the resource allocation scheme vector population; advanced recombination operators and single-point crossover operators are used for crossover processing; In the step 311, mutation; The mutation is to generate a new row resource configuration scheme vector by replacing a part of a row resource configuration scheme vector in the resource configuration scheme vector population with other parts of the row resource configuration scheme vector; a discrete mutation operator is used to perform the mutation operation; In the step 312, reinsert; Fitness-based reinsertion is used to ensure that P s ×(1-GGAP) most adapted split vectors are always propagated continuously to the next generation; After executing step 312, go to step 305 for execution; Steps 305 to 312 are repeated until a termination condition is met.

Images