A Scheduling Optimization Method for Ship Maintenance Task Scheduling Problem

By combining local precise algorithms and new neighborhood structure scheduling optimization methods, the problems of distributed maintenance and deterioration effects in ship maintenance scheduling are solved, efficient maintenance task allocation and scheduling are achieved, and service efficiency and algorithms of ship manufacturing enterprises are improved. performance.

CN114936713BActive Publication Date: 2024-11-22HEFEI UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210693091.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-06-17
Publication Date
2024-11-22
Estimated Expiration
2042-06-17

AI Technical Summary

Technical Problem

The existing technology is difficult to effectively solve the problem of ship maintenance scheduling, especially in distributed maintenance environments. Traditional scheduling algorithms ignore the deterioration effect and distributed characteristics of maintenance tasks, resulting in inefficient scheduling schemes and it is difficult to complete maintenance tasks within a reasonable time. .

Method used

A scheduling optimization method combining local precision algorithms and a variety of new neighborhood structures was designed. Through integer encoding, heuristic algorithms, vibration operation and lower bound solution algorithms, the allocation and scheduling of ship maintenance tasks were optimized, and the deterioration of maintenance tasks was considered. and distributed characteristics improve the algorithm's solution efficiency and flexibility.

Benefits of technology

This method can give an approximate optimal solution to the ship maintenance task within a reasonable time, shorten the maintenance service span, improve the service efficiency of ship manufacturing enterprises, and improve the algorithm performance through a variety of heuristics and precise algorithms, which can be used to achieve multinomials Calculate the lower bound of the problem in time.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114936713B_ABST
    Figure CN114936713B_ABST
Patent Text Reader

Abstract

The present invention discloses a scheduling optimization method for ship maintenance task scheduling problems, and belongs to the technical field of ship maintenance scheduling. The present invention sorts workpieces according to different execution times of maintenance tasks, distances from spare parts warehouses, and deterioration characteristics, and assigns maintenance tasks to different maintenance personnel, with the goal of minimizing the maintenance service span, thereby improving service efficiency and reducing service costs for shipbuilding enterprises; and the problem lower bound can be calculated in polynomial time, which can be used to judge the pros and cons of scheduling plans, as an important evaluation criterion for the quality of scheduling plans.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of ship maintenance scheduling, and in particular to a scheduling optimization method for ship maintenance task scheduling problems. Background Art

[0002] The scheduling of ship maintenance tasks has the following main characteristics: First, the user's ship often goes to the corresponding port to wait for maintenance before it needs maintenance. These ports are distributed in different geographical locations, which leads to the dispersion of maintenance tasks. Maintenance personnel need to carry the required parts and go back and forth between various ports and spare parts centers to complete the maintenance tasks; secondly, the maintenance tasks are deteriorating, that is, when the ship waits too long in the port, the time required for its corresponding maintenance tasks will be extended. This is due to the deterioration of the mechanical system caused by untimely maintenance, which increases the difficulty of executing the maintenance tasks. Due to the above characteristics of the ship maintenance scheduling problem, many traditional workshop scheduling algorithms cannot effectively give reasonable scheduling plans, which is not conducive to the further development and competitiveness of shipbuilding enterprises.

[0003] Variable Neighborhood Search (VNS) is a and Hansen proposed a meta-heuristic algorithm. The main idea is to lock the global optimal solution through systematic changes in the neighborhood. The algorithm has the characteristics of simple structure and easy implementation. Behnamian et al. used a variable neighborhood search algorithm to solve a type of parallel machine scheduling problem considering the setup time. Diana et al. designed a method based on variable neighborhood search to solve a type of unrelated parallel machine robust scheduling optimization problem. However, there are also some problems in the application of the variable neighborhood search algorithm. The neighborhood structure provided by the basic variable neighborhood search algorithm is relatively simple. It has high efficiency in solving some simple problems, but it is often easy to fall into local optimality when solving complex ship scheduling problems. It is necessary to design a reasonable neighborhood structure according to the characteristics of the ship maintenance scheduling problem and combine some local precise algorithms to give a reliable approximate optimal solution within a reasonable time.

[0004] The existing scheduling optimization method for ship maintenance task scheduling problem has the following steps:

[0005] (1) At the problem level, traditional scheduling literature often assumes that the time of production tasks is fixed, ignoring the deterioration effect of maintenance tasks and failing to consider that the maintenance task time will change as the waiting time increases. In addition, traditional scheduling takes place within the workshop and fails to propose effective solutions based on the characteristics of distributed maintenance tasks. Some supply chain scheduling problems consider product delivery but do not consider the alternation between maintenance tasks and transportation processes. The structure of the problem does not conform to the typical characteristics of ship maintenance scheduling problems, and it is necessary to make effective scheduling arrangements based on the characteristics of ship maintenance scheduling problems.

[0006] 2) At the method level, the variable neighborhood search algorithm is an effective intelligent algorithm that has been well applied and verified in various scheduling problems. However, the variable neighborhood search algorithm itself has a relatively simple neighborhood design. When solving some large-scale complex problems, more effective neighborhood structures are needed to ensure algorithm performance.

[0007] In order to solve the above problems, a scheduling optimization method for ship maintenance task scheduling problem is proposed. Summary of the invention

[0008] The technical problem to be solved by the present invention is: how to solve the problem of distributed ship maintenance scheduling, and use limited maintenance personnel to complete the maintenance scheduling problem distributed in different ports in the shortest time. A scheduling optimization method for ship maintenance task scheduling problem is provided. According to the characteristics of the distributed ship maintenance scheduling problem, the method designs a local precise algorithm and a variety of new neighborhood structures, which can overcome the shortcomings of the classic variable neighborhood search algorithm such as slow convergence speed and insufficient convergence, and further improve the solution efficiency of the algorithm.

[0009] The present invention solves the above technical problems through the following technical solutions, and the present invention comprises the following steps:

[0010] S1: Initialize parameters xmax, itmax, and encode the problem using integer encoding. The solution to the problem is represented by V = {v1, v2, …, v n},v i Indicates that the i-th maintenance task is assigned to the v-th i Maintenance personnel, 1≤v i ≤m, where xmax is the preset maximum number of Shaking vibrations, itmax is the preset maximum running time of the algorithm, and m represents the total number of maintenance personnel;

[0011] S2: Set it = 0, x = 1, and x′ = 1; execute the heuristic algorithm to obtain the initial solution V 0 , calculate V 0 The service span C of the corresponding scheduling scheme max (V 0), the service span is the maximum completion time;

[0012] S3: V 0 Perform the vibration operation Shaking(it,itmax,x) to get V 1 ;

[0013] S4: V 1 Execute VND x′ Method to get V 2 , calculate V 2 The service span C of the corresponding scheduling scheme max (V 2 );

[0014] S5: Determine the service span C max (V 0 ) and C max (V 2 ) is related to the size of the max (V 2 ) <C max (V 0 ), then set x = 1 and set V 2 Assign to V 0 , otherwise set x=x+1 and x′=x′+1;

[0015] S6: Determine the size relationship between x and xmax. If x>xmax, set x=1. Otherwise, do not adjust x and directly execute step S7.

[0016] S7: Determine whether x′ is greater than 4. If x′>4, set x′=1. Otherwise, do not adjust x′ and directly execute step S8.

[0017] S8: Set the current algorithm running time it = CPUtime(), and determine the size relationship between it and itmax. If it≤itmax, return to step S3, otherwise directly execute step S9, where CPUtime() represents the time the current algorithm has been running;

[0018] S9: Execute the lower bound solution algorithm and output V 0 And the lower bound LB of the problem.

[0019] Furthermore, in step S2 or step S4, the scheduling scheme service span calculation and scheduling mechanism specifically includes the following steps:

[0020] S201: The maintenance tasks are encoded according to V = {v1, v2, ..., v n}Assigned to maintenance personnel;

[0021] S202: For each maintenance task of each maintenance personnel, calculate the ranking value f[j],k =d [j],k +2T [j],k +bT [j],k , where d [j],k , represents the basic time required for the jth maintenance task faced by the kth maintenance personnel, T [j],k represents the time required for the kth maintenance personnel to complete the jth maintenance task at the spare parts center;

[0022] S203: For each maintenance personnel, the maintenance task assigned to him is calculated according to f [j],k =d [j],k +2T [j],k +bT [j],k Sort in non-decreasing order, and then calculate the completion time of each maintenance worker

[0023] S204: Output C max (V) = max{C k} and the order in which each maintenance personnel performs maintenance tasks, where max{C k} means take C k The maximum value among 1≤k≤m.

[0024] Furthermore, in step S2, the heuristic algorithm specifically includes the following steps:

[0025] S211: All maintenance tasks are carried out according to f i =d i +2T i +bT i Non-decreasing sort of ;

[0026] S212: set i=1;

[0027] S213: Assign the i-th workpiece to the maintenance personnel with the shortest current completion time, and then execute the scheduling scheme service span calculation and scheduling mechanism in step S2 or S4 to update the completion time C of each maintenance personnel. k ;

[0028] S214: Set i=i+1. If i≤n, where n represents the total number of ship maintenance tasks, return to step S213. Otherwise, output the maintenance task allocation plan and end the operation.

[0029] Furthermore, in step S3, the vibration operation Shaking(it,itmax,x) comprises the following steps:

[0030] S31: set x1=1;

[0031] S32: Generate a random number x2 between 0 and 1;

[0032] S33: If The solution vector V for the input 0 Execute RTNS to get the new solution vector V 1 ;

[0033] S34: If The solution vector V for the input 0 Execute RINS to get the new solution vector V 1 ;

[0034] S35: If The solution vector V for the input 0 Execute RSNS to get the new solution vector V 1 ;

[0035] S36: Set x1=x1+1; if x1≤x, execute step S32; otherwise, output V 1 And end the run.

[0036] Furthermore, in step S4, when x′=1, VND x′ The method specifically comprises the following steps:

[0037] S401: Set x1=1;

[0038] S402: If x1=1, input solution vector V 1 Execute the flip operator RTNS to get the new solution vector V 2 ;

[0039] S403: If x1=2, input solution vector V 1 Execute the insertion operator RINS to obtain the new solution vector V 2 ;

[0040] S404: If x1=3, input solution vector V 1 Execute the exchange operator RSNS to get the new solution vector V 2 ;

[0041] S405: Calculate C using the scheduling scheme service span calculation and scheduling mechanism in step S2 or S4 max (V 2 ), if C max (V 2 ) <C max (V 1 ), set x1 = 1 and set V 2 Assign to V 1 , otherwise, let x1=x1+1;

[0042] S406: If x1≤3, return to step S402, otherwise, output V and end the operation;

[0043] When x′=2, VND x′ The method specifically comprises the following steps:

[0044] S411: Set x1=1;

[0045] S412: If x1=1, the input solution vector V 1 Execute the insertion operator RINS to obtain the new solution vector V 2 ;

[0046] S413: If x1=2, the input solution vector V 1 Execute the flip operator RTNS to get the new solution vector V 2 ;

[0047] S414: If x1=3, the input solution vector V 1 Execute the exchange operator RSNS to get the new solution vector V 2 ;

[0048] S415: Calculate C using the scheduling scheme service span calculation and scheduling mechanism in step S2 or S4 max (V 2 ), if C max (V 2 ) <C max (V 1 ), set x1 = 1 and set V 2 Assign to V 1 , otherwise, let x1=x1+1;

[0049] S416: If x1≤3, return to step S412; otherwise, output V 1 And end the operation;

[0050] When x′=3, VND x′ The method specifically comprises the following steps:

[0051] S421: Set x1=1;

[0052] S422: If x1=1, the input solution vector V 1 Execute the flip operator RTNS to get the new solution vector V 2 ;

[0053] S423: If x1=2, execute the exchange operator RSNS on the input solution vector V to obtain a new solution vector V 2 ;

[0054] S424: If x1=3, perform the insertion operator RINS on the input solution vector V to obtain a new solution vector V 2 ;

[0055] S425: Calculate C using the scheduling scheme service span calculation and scheduling mechanism in step S2 or S4 max (V 2 ), if C max (V 2 ) <C max (V 1 ), set x1 = 1 and set V 2 Assign to V 1 , otherwise, let x1=x1+1;

[0056] S426: If x1≤3, return to step S422; otherwise, output V 1 And end the operation;

[0057] When x′=4, VND x′ The method specifically comprises the following steps:

[0058] S431: Set x1=1;

[0059] S432: If x1=1, the input solution vector V 1 Execute the exchange operator RSNS to get the new solution vector V 2 ;

[0060] S433: If x1=2, the input solution vector V 1 Execute the insertion operator RINS to obtain the new solution vector V 2 ;

[0061] S434: If x1=3, the input solution vector V 1 Execute the flip operator RTNS to get the new solution vector V 2 ;

[0062] S435: Calculate C using the scheduling scheme service span calculation and scheduling mechanism in step S2 or S4 max (V 2 ), if C max (V 2 ) <C max (V 1 ), set x1 = 1 and set V 2 Assign to V 1 , otherwise, let x1=x1+1;

[0063] S436: If x1≤3, return to step S432; otherwise, output V 1 And end the run.

[0064] Furthermore, the execution process of the flip operator RTNS includes the following steps:

[0065] A1: Randomly generate random numbers x1 and x2 between 1 and n, where x1+1 <x2;

[0066] A2: For the input solution vector V 1 ={v1,v2,…,v n},Will and The sequence between them is reversed and rearranged, and a new solution vector V is output 2 .

[0067] Furthermore, the execution process of the insertion operator RINS includes the following steps:

[0068] B1: Randomly generate random numbers x1 and x2 between 1 and n, where x1 <x2;

[0069] B2: For the input solution vector V 1 ={v1,v2,…,v n},Will Insert into Before, we get the rearranged V 1 , output the new solution vector V 2 .

[0070] Furthermore, the execution process of the exchange operator RSNS includes the following steps:

[0071] C1: Randomly generate random numbers x1 and x2 between 1 and n, where x1 <x2;

[0072] C2: Input solution vector V 1 ={v1,v2,…,v n}middle and The values ​​of are swapped and rearranged, and the new solution vector V is output 2 .

[0073] Furthermore, in step S9, the lower bound solving algorithm specifically includes the following steps:

[0074] S91: Set k=1 and i=1;

[0075] S92: For each maintenance task, calculate f i =d i +2T i +bT i , where T i represents the travel time from the ith maintenance task to the spare parts warehouse, d i represents the basic time required for the ith maintenance task, and b represents the deterioration coefficient of the maintenance task;

[0076] S93: The maintenance task is carried out according to fi Reorder in non-decreasing order;

[0077] S94: Assign the i-th maintenance task to the k-th maintenance personnel, setting k=k+1 and i=i+1;

[0078] S95: If k≤m and i≤n, return to step 4, otherwise execute step S96;

[0079] S96: If k>m and i≤n, set k=1 and return to step S94, otherwise execute step S97;

[0080] S97: If i>n, calculate Output lower bound Indicates not greater than The maximum integer.

[0081] Compared with the prior art, the present invention has the following advantages: the scheduling optimization method for ship maintenance task scheduling problems is based on actual ship maintenance optimization scenarios, takes into account the deterioration and distributed characteristics of maintenance tasks, and designs corresponding approximate algorithms and precise algorithms in combination with the characteristics of multiple maintenance tasks and multiple maintenance personnel in the maintenance scheduling process. It can not only ensure that the optimal solution is obtained in some local optimization problems, but also quickly obtain the global approximate optimal solution of the entire problem. It can be used in the maintenance service scheduling process of shipbuilding enterprises, shorten the maintenance service span of shipbuilding enterprises, and improve the service efficiency of shipbuilding enterprises; at the same time, it integrates a variety of heuristic algorithms, precise algorithms and neighborhood structures, proposes a new local sorting scheduling algorithm, innovates the vibration Shaking operation in the variable neighborhood search algorithm, improves the flexibility of the Shaking operation, can effectively improve the algorithm performance, and can calculate the problem lower bound in polynomial time, which can be used as an important evaluation criterion for the quality of scheduling solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 It is a schematic diagram of the architecture of the ship maintenance scheduling problem involved in the present invention. DETAILED DESCRIPTION

[0083] The following is a detailed description of an embodiment of the present invention. This embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation method and a specific operation process are given, but the protection scope of the present invention is not limited to the following embodiment.

[0084] This embodiment provides a technical solution: a scheduling optimization method for ship maintenance task scheduling problem, which is to design an intelligent optimization and lower bound solving algorithm for ship maintenance scheduling problem, and give a near-optimal ship maintenance scheduling solution and its lower bound within a reasonable time. In a ship maintenance scheduling case, there are n ship maintenance tasks and m maintenance personnel. The ship maintenance tasks are initially moored at different ports, and the ship maintenance personnel are initially in the spare parts center. i represents the i-th maintenance task, M k represents the kth maintenance personnel, T i represents the travel time from the ith maintenance task to the spare parts warehouse, d i represents the basic time required for the ith maintenance task, b represents the deterioration coefficient of the maintenance task (the deterioration coefficient is obtained from historical data and experience), C k represents the completion time of the kth maintenance personnel; C max represents the maximum completion time, i.e., the service span. The goal of the problem is to minimize the service span C max The actual time of each maintenance task is expressed as ( represents the actual time, A is actual; t represents the start time of the maintenance task), that is, as the start time is delayed, the actual time of the maintenance task will continue to extend under the influence of the deterioration effect.

[0085] The specific steps of the ship maintenance scheduling optimization and lower bound solving method in this embodiment are as follows:

[0086] (I) Initialize the parameters xmax, itmax (xmax is the preset maximum number of shaking vibrations, itmax is the preset maximum running time of the algorithm), and encode the problem using integer encoding. The solution to the problem can be expressed as V = {v1, v2, …, v n},v i Indicates that the i-th maintenance task is assigned to the v-th i Maintenance personnel, 1≤v i ≤m, where m represents the total number of maintenance personnel;

[0087] (ii) Set it = 0, x = 1, and x′ = 1; execute the heuristic algorithm to obtain the initial solution V 0 , calculate V 0 The service span C of the corresponding scheduling scheme max (V 0 );

[0088] (III) V 0 Perform the vibration operation Shaking(it,itmax,x) to get V 1 ;

[0089] (IV) V 1 Execute VND x′ Method to get V 2 , calculate V 2 The service span C of the corresponding scheduling scheme max (V 2 );

[0090] (V) Determining the service span C max (V 0 ) and C max (V 2 ) is related to the size of the max (V 2 ) <C max (V 0 ), then set x = 1 and set V 2 Assign to V 0 , otherwise set x=x+1 and x′=x′+1;

[0091] (vi) Determine the size relationship between x and xmax. If x>xmax, set x=1. Otherwise, do not adjust x and directly execute step 7.

[0092] (VII) Determine whether x′ is greater than 4. If x′>4, set x′=1. Otherwise, do not adjust x′ and directly execute step 8.

[0093] (VIII) Set the current algorithm running time it = CPUtime() (CPUtime() means the time the current algorithm has been running), determine the size relationship between it and itmax, if it≤itmax, return to step 3, otherwise directly execute step 9, where;

[0094] (IX) Execute the lower bound solution algorithm and output V 0 And the lower bound LB of the problem ends the algorithm.

[0095] In this embodiment, the scheduling scheme service span calculation and scheduling mechanism mentioned in the above steps 2 and 4 specifically include the following steps:

[0096] Step 1: Encode the maintenance tasks according to V = {v1, v2, …, v n}Assigned to maintenance personnel;

[0097] Step 2: For each maintenance task of each maintenance worker, calculate the ranking value f [j],k =d [j],k +2T [j],k +bT [j],k , where d [j],k , represents the basic time required for the jth maintenance task faced by the kth maintenance personnel, T [j],krepresents the time required for the kth maintenance personnel to complete the jth maintenance task at the spare parts center;

[0098] Step 3: For each maintenance personnel, assign the maintenance tasks to them according to f [j],k =d [j],k +2T [j],k +bT [j],k Sort in non-decreasing order, and then calculate the completion time of each maintenance worker

[0099] Step 4: Output C max (V) = max{C k} and the order in which each maintenance personnel performs maintenance tasks, where max{C k} means take C k The maximum value among 1≤k≤m.

[0100] In this embodiment, the heuristic algorithm in step 2 specifically includes the following steps:

[0101] Step 1: All maintenance tasks are carried out according to the i =d i +2T i +bT i Non-decreasing sort of ;

[0102] Step 2: Set i=1;

[0103] Step 3: Assign the i-th workpiece to the maintenance worker with the shortest current completion time, and then execute the scheduling scheme service span calculation and scheduling mechanism mentioned in steps 2 and 4 to update the completion time C of each maintenance worker. k ;

[0104] Step 4: Set i=i+1. If i≤n, where n represents the total number of ship maintenance tasks, return to step 3. Otherwise, output the maintenance task allocation plan and end the algorithm.

[0105] In this embodiment, the non-decreasing sorting means that the maintenance tasks are arranged from small to large according to the basic required time.

[0106] In this embodiment, the Shaking(it, itmax, x) method in step 3 includes the following steps:

[0107] Step 1: Set x1=1;

[0108] Step 2: Generate a random number x2 between 0 and 1;

[0109] Step 3: If The input solution vector V (i.e. V 0 ) executes RTNS to obtain a new solution vector V′ (i.e. V1 );

[0110] Step 4: If Execute RINS on the input solution vector V to obtain a new solution vector V′;

[0111] Step 5: If Perform RSNS on the input solution vector V to obtain a new solution vector V′;

[0112] Step 6: Set x1=x1+1; if x1≤x, execute step 2, otherwise, output V′ and end the algorithm.

[0113] In this embodiment, the VND in step 4 is x′ The method specifically comprises the following steps:

[0114] The order of executing the neighborhood for VND1, VND2, VND3, and VND4 is RTNS→RINS→RSNS, RINS→RTNS→RSNS, RTNS→RSNS→RINS, and RSNS→RINS→RTNS. The other steps are the same. The following takes VND1 as an example to introduce VND x′ The steps of the method;

[0115] Step 1: Set x1=1;

[0116] Step 2: If x1=1, input the solution vector V (i.e. V 1 ) executes RTNS to obtain a new solution vector V′ (i.e. V 2 );

[0117] Step 3: If x1=2, perform RINS on the input solution vector V to obtain a new solution vector V′;

[0118] Step 4: If x1=3, perform RSNS on the input solution vector V to obtain a new solution vector V′;

[0119] Step 5: Calculate C using the scheduling scheme service span calculation and scheduling mechanism mentioned in steps 2 and 4 max (V′), if C max (V′) <C max (V), set x1 = 1 and assign V′ to V, otherwise, set x1 = x1 + 1;

[0120] Step 6: If x1≤3, return to step 2, otherwise, output V and end the algorithm;

[0121] In this embodiment, the execution process of the neighborhood operation operators RTNS, RINS, and RSNS includes the following steps:

[0122] Flip operator RTNS:

[0123] Step 1: Randomly generate random numbers x1 and x2 between 1 and n, where x1+1 <x2;

[0124] Step 2: For the input solution vector V = {v1, v2, …, v n},Will and The sequence between them is reversed and rearranged, and a new solution vector V is output;

[0125] Insert operator RINS:

[0126] Step 1: Randomly generate random numbers x1 and x2 between 1 and n, where x1 <x2;

[0127] Step 2: For the input solution vector V = {v1, v2, …, v n},Will Insert into Before, get the rearranged V and output the new solution vector V;

[0128] Exchange operator RSNS:

[0129] Step 1: Randomly generate random numbers x1 and x2 between 1 and n, where x1 <x2;

[0130] Step 2: Input solution vector V = {v1, v2, …, v n}middle and The values ​​of are swapped and rearranged, and a new solution vector V is output.

[0131] In this embodiment, the lower bound solving algorithm in step nine is:

[0132] Step 1: Set k=1 and i=1;

[0133] Step 2: For each maintenance task, calculate f i =d i +2T i +bT i ;

[0134] Step 3: Assign maintenance tasks to i Reorder in non-decreasing order;

[0135] Step 4: Assign the i-th maintenance task to the k-th maintenance personnel, setting k=k+1 and i=i+1;

[0136] Step 5: If k≤m and i≤n, return to step 4, otherwise go to step 6;

[0137] Step 6: If k>m and i≤n, set k=1 and return to step 4, otherwise go to step 7;

[0138] Step 7: If i>n, calculate Output lower bound Indicates not greater than The maximum integer.

[0139] It should be noted that the above lower bound solving algorithm obtains the lower bound of the problem. In practical applications, the closer the result of the scheduling optimization calculation is to the lower bound, the better the result is, and the farther it is from the lower bound, the worse the result is.

[0140] In summary, the scheduling optimization method for ship maintenance task scheduling problems in the above-mentioned embodiment is based on the actual ship maintenance optimization scenario, takes into account the deterioration and distributed characteristics of maintenance tasks, and designs corresponding approximate algorithms and precise algorithms in combination with the characteristics of multiple maintenance tasks and multiple maintenance personnel in the maintenance scheduling process. It can not only ensure that the optimal solution is obtained in some local optimization problems, but also quickly obtain the global approximate optimal solution of the entire problem. It can be used in the maintenance service scheduling process of shipbuilding enterprises, shorten the maintenance service span of shipbuilding enterprises, and improve the service efficiency of shipbuilding enterprises; at the same time, it integrates a variety of heuristic algorithms, precise algorithms and neighborhood structures, proposes a new local sorting scheduling algorithm, innovates the vibration Shaking operation in the variable neighborhood search algorithm, improves the flexibility of the Shaking operation, can effectively improve the algorithm performance, and can calculate the problem lower bound in polynomial time, which can be used as an important evaluation criterion for the quality of scheduling solutions.

[0141] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A scheduling optimization method for ship maintenance task scheduling problem, characterized in that: The following steps are involved: S1: Initialize parameters xmax, itmax, and encode the problem using integer encoding. The solution to the problem is represented by V = {v1, v2, …, v n },v i Indicates that the i-th maintenance task is assigned to the v-th i Maintenance personnel, 1≤v i ≤m, where xmax is the preset maximum number of Shaking vibrations, itmax is the preset maximum running time of the algorithm, and m represents the total number of maintenance personnel; S2: Set it = 0, x = 1, and x′ = 1; execute the heuristic algorithm to obtain the initial solution V 0 , calculate V 0 The service span C of the corresponding scheduling scheme max (V 0 ), the service span is the maximum completion time; S3: V 0 Perform the vibration operation Shaking(it,itmax,x) to get V 1 ; S4: V 1 Execute VND x′ Method to get V 2 , calculate V 2 The service span C of the corresponding scheduling scheme max (V 2 ); S5: Determine the service span C max (V 0 ) and C max (V 2 ) is related to the size of the max (V 2 ) <C max (V 0 ), then set x = 1 and set V 2 Assign to V 0 , otherwise set x=x+1 and x′=x′+1; S6: Determine the size relationship between x and xmax. If x>xmax, set x=1. Otherwise, do not adjust x and directly execute step S7. S7: Determine whether x′ is greater than 4. If x′>4, set x′=1. Otherwise, do not adjust x′ and directly execute step S8. S8: Set the current algorithm running time it = CPUtime(), and determine the size relationship between it and itmax. If it≤itmax, return to step S3, otherwise directly execute step S9, where CPUtime() represents the time the current algorithm has been running; S9: Execute the lower bound solution algorithm and output V 0 And the lower bound LB of the problem; In step S2 or step S4, the scheduling scheme service span calculation and scheduling mechanism specifically includes the following steps: S201: The maintenance tasks are encoded according to V = {v1, v2, ..., v n }Assigned to maintenance personnel; S202: For each maintenance task of each maintenance personnel, calculate the ranking value f [j],k =d [j],k +2T [j],k +bT [j],k , where d [j],k , represents the basic time required for the jth maintenance task faced by the kth maintenance personnel, T [j],k represents the time required for the kth maintenance personnel to reach the spare parts center for the jth maintenance task, and b represents the deterioration coefficient of the maintenance task; S203: For each maintenance personnel, the maintenance task assigned to him is calculated according to f [j],k =d [j],k +2T [j],k +bT [j],k Sort in non-decreasing order, and then calculate the completion time of each maintenance worker Where n represents the total number of ship maintenance tasks; S204: Output C max (V) = max{C k } and the order in which each maintenance personnel performs maintenance tasks, where max{C k } means take C k The maximum value among 1≤k≤m.

2. The scheduling optimization method for ship maintenance task scheduling problem according to claim 1 is characterized by: In step S2, the heuristic algorithm specifically includes the following steps: S211: All maintenance tasks are carried out according to f i =d i +2T i +bT i Non-decreasing sort of ; S212: set i=1; S213: Assign the i-th workpiece to the maintenance personnel with the shortest current completion time, and then execute the scheduling scheme service span calculation and scheduling mechanism in step S2 or S4 to update the completion time C of each maintenance personnel. k ; S214: Set i=i+1. If i≤n, where n represents the total number of ship maintenance tasks, return to step S213. Otherwise, output the maintenance task allocation plan and end the operation.

3. The scheduling optimization method for ship maintenance task scheduling problem according to claim 2 is characterized by: In step S3, the vibration operation Shaking(it,itmax,x) comprises the following steps: S31: set x1=1; S32: Generate a random number x2 between 0 and 1; S33: If The solution vector V for the input 0 Execute RTNS to get the new solution vector V 1 ; S34: If The solution vector V for the input 0 Execute RINS to get the new solution vector V 1 ; S35: If The solution vector V for the input 0 Execute RSNS to get the new solution vector V 1 ; S36: Set x1=x1+1; if x1≤x, execute step S32; otherwise, output V 1 And end the run.

4. The scheduling optimization method for ship maintenance task scheduling problem according to claim 3 is characterized by: In step S4, when x ′ =1, VND x′ The method specifically comprises the following steps: S401: Set x1=1; S402: If x1=1, input solution vector V 1 Execute the flip operator RTNS to get the new solution vector V 2 ; S403: If x1=2, input solution vector V 1 Execute the insertion operator RINS to obtain the new solution vector V 2 ; S404: If x1=3, input solution vector V 1 Execute the exchange operator RSNS to get the new solution vector V 2 ; S405: Calculate C using the scheduling scheme service span calculation and scheduling mechanism in step S2 or S4 max (V 2 ), if C max (V 2 ) <C max (V 1 ), set x1 = 1 and set V 2 Assign to V 1 , otherwise, let x1=x1+1; S406: If x1≤3, return to step S402, otherwise, output V and end the operation; When x′=2, VND x′ The method specifically comprises the following steps: S411: Set x1=1; S412: If x1=1, the input solution vector V 1 Execute the insertion operator RINS to obtain the new solution vector V 2 ; S413: If x1=2, the input solution vector V 1 Execute the flip operator RTNS to get the new solution vector V 2 ; S414: If x1=3, the input solution vector V 1 Execute the exchange operator RSNS to get the new solution vector V 2 ; S415: Calculate C using the scheduling scheme service span calculation and scheduling mechanism in step S2 or S4 max (V 2 ), if C max (V 2 ) <C max (V 1 ), set x1 = 1 and set V 2 Assign to V 1 , otherwise, let x1=x1+1; S416: If x1≤3, return to step S412; otherwise, output V 1 And end the operation; When x′=3, VND x′ The method specifically comprises the following steps: S421: Set x1=1; S422: If x1=1, the input solution vector V 1 Execute the flip operator RTNS to get the new solution vector V 2 ; S423: If x1=2, execute the exchange operator RSNS on the input solution vector V to obtain a new solution vector V 2 ; S424: If x1=3, perform the insertion operator RINS on the input solution vector V to obtain a new solution vector V 2 ; S425: Calculate C using the scheduling scheme service span calculation and scheduling mechanism in step S2 or S4 max (V 2 ), if C max (V 2 ) <C max (V 1 ), set x1 = 1 and set V 2 Assign to V 1 , otherwise, let x1=x1+1; S426: If x1≤3, return to step S422; otherwise, output V 1 And end the operation; When x′=4, VND x′ The method specifically comprises the following steps: S431: Set x1=1; S432: If x1=1, the input solution vector V 1 Execute the exchange operator RSNS to get the new solution vector V 2 ; S433: If x1=2, the input solution vector V 1 Execute the insertion operator RINS to obtain the new solution vector V 2 ; S434: If x1=3, the input solution vector V 1 Execute the flip operator RTNS to get the new solution vector V 2 ; S435: Calculate the service span of the scheduling scheme and the scheduling mechanism in step S2 or S4 to calculate C max (V 2 ), if C max (V 2 ) <C max (V 1 ), set x1 = 1 and set V 2 Assign to V 1 , otherwise, let x1=x1+1; S436: If x1≤3, return to step S432; otherwise, output V 1 And end the run.

5. The scheduling optimization method for ship maintenance task scheduling problem according to claim 4 is characterized by: The execution process of the flip operator RTNS includes the following steps: A1: Randomly generate random numbers x1 and x2 between 1 and n, where x1+1 <x2; A2: For the input solution vector V 1 ={v1,v2,…,v n }, and v x1 and v x2 The sequence between them is reversed and rearranged, and a new solution vector V is output 2 .

6. The scheduling optimization method for ship maintenance task scheduling problem according to claim 4 is characterized by: The execution process of the insertion operator RINS includes the following steps: B1: Randomly generate random numbers x1 and x2 between 1 and n, where x1 <x2; B2: For the input solution vector V 1 ={v1,v2,…,v n }, and change v x2 Insert into v x1 Before, we get the rearranged V 1 , output the new solution vector V 2 .

7. The scheduling optimization method for ship maintenance task scheduling problem according to claim 4 is characterized by: The execution process of the exchange operator RSNS includes the following steps: C1: Randomly generate random numbers x1 and x2 between 1 and n, where x1 <x2; C2: Input solution vector V 1 ={v1,v2,…,v n } x1 and v x2 The values ​​of are swapped and rearranged, and the new solution vector V is output 2 .

8. The scheduling optimization method for ship maintenance task scheduling problem according to claim 4 is characterized by: In step S9, the lower bound solving algorithm specifically includes the following steps: S91: Set k=1 and i=1; S92: For each maintenance task, calculate f i =d i +2T i +bT i , where T i represents the travel time from the ith maintenance task to the spare parts warehouse, d i represents the basic time required for the ith maintenance task, and b represents the deterioration coefficient of the maintenance task; S93: The maintenance task is carried out according to f i Reorder in non-decreasing order; S94: Assign the i-th maintenance task to the k-th maintenance personnel, setting k=k+1 and i=i+1; S95: If k≤m and i≤n, return to step 4, otherwise execute step S96; S96: If k>m and i≤n, set k=1 and return to step S94, otherwise execute step S97; S97: If i>n, calculate Output lower bound Indicates not greater than The maximum integer.