Consider the high-end equipment production rescheduling optimization method under machine failure

Abstract

The application provides a high-end equipment production rescheduling optimization method considering machine failure, and relates to the technical field of task scheduling. In the application, input parameters are set based on workpiece data, machine data and rescheduling data in the production process of a high-end equipment workshop; after machine failure occurs, a workpiece set, a machine set and a workpiece processing start time are updated; based on the updated input parameters, a hybrid particle swarm algorithm based on neighborhood search is used to completely reschedule the workpieces, a global optimal solution is obtained, and corresponding starting spare machine quantity, a production rescheduling scheme, a rescheduling production cost, a rescheduling spare machine starting cost and a rescheduling conflict cost are decoded. The optimization objectives of rescheduling processing time and cost are comprehensively considered, and a workpiece sequence scheduling model of a manufacturing interruption flexible workshop is constructed. The improved particle swarm algorithm can effectively reduce the production manufacturing cost and enhance the ability of enterprises to deal with sudden production interruption problems.

Description

Technical Field

[0001] This invention relates to the field of task scheduling technology, and specifically to a method, system, storage medium, and electronic device for optimizing the rescheduling of high-end equipment production under machine failure conditions. Background Technology

[0002] The production environment for high-end equipment is generally complex, and various factors such as natural disasters, urgent orders, and equipment failures can lead to production interruptions. Interruptions can render existing workshop scheduling plans inapplicable, forcing companies to readjust their scheduling arrangements.

[0003] In high-end equipment manufacturing enterprises, equipment failures frequently occur in production workshops. Downtime caused by equipment failures prevents the processing of workpieces originally scheduled for that equipment, and the remaining working machines may be unable to complete the workpieces within the delivery deadline, impacting the enterprise's production schedule and potentially causing losses. To mitigate these situations, enterprises typically reschedule production plans. By comprehensively deciding on the number of machines deployed and the order of workpiece processing, enterprises ensure on-time completion of workpieces and reduce the overall cost of rescheduling. Currently, research on the rescheduling problem in high-end equipment production due to equipment failures is relatively limited. Given the manufacturing enterprises' pursuit of high efficiency and low energy consumption, research on this rescheduling problem has significant theoretical and practical value. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] To address the shortcomings of existing technologies, this invention provides a method, system, storage medium, and electronic device for optimizing the rescheduling of high-end equipment production under machine failure scenarios, thus solving the technical problem of rescheduling workshops with backup machines.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A method for optimizing the rescheduling of high-end equipment production under machine failure scenarios includes:

[0009] Based on workpiece data, machine data and rescheduling data during the production process in the high-end equipment workshop, set input parameters, including at least the workpiece set, machine set, workpiece processing start and end time, standby machine start-up cost, conflict cost, machine failure time point and fixed machine maintenance time.

[0010] After a machine failure occurs, update the workpiece set, machine set, and workpiece processing start time;

[0011] Based on the updated input parameters, a hybrid particle swarm optimization algorithm based on neighborhood search is used to completely reschedule the workpieces, obtain the global optimal solution, and decode the corresponding number of standby machines to start, production rescheduling scheme, production cost of rescheduling, cost of starting standby machines for rescheduling, and cost of rescheduling conflicts.

[0012] Preferably, the input parameters specifically include:

[0013] The set of n initial workpieces, J = [J1, J2, ..., Jn], consists of a workpiece set J = [J1, J2, ..., Jn]. n ];

[0014] A set of workpieces consisting of l unprocessed workpieces

[0015] The set of machines MM = {MM1, MM2, ..., MM} u BM u+1 BM u+2 , ..., BM p}, where MM j For the machines that perform the initial production operation, u is the number of machines for the initial production operation, BM j p represents the number of newly activated standby machines to be rescheduled.

[0016] Assume CM j CM (China Machine Tool) q A fault occurs, q∈[1,2,...,u];

[0017] CM q The time interval t from the start of processing to the occurrence of failure follows an exponential distribution, and the failed machine CM q The maintenance time d follows an exponential distribution;

[0018] Machine CM up to time t j The set of processed workpieces F j The set of all workpieces processed on all machines up to time t;

[0019] Non-faulty machine CM at time t j The set of workpieces being processed is Z. j Z is the set of all workpieces being processed on all non-faulty machines at time t;

[0020] P ij For workpiece J i In Machine MM j The time spent processing on top, C i Indicates the initial scheduling of workpiece J i Processing completion time Indicates rescheduling of workpiece J i Processing completion time; Sj For rescheduled machine MM j Start time of the first workpiece processing;

[0021] c ij For workpiece J i In Machine MM j The unit time processing cost;

[0022] ε BM Startup cost for each backup machine;

[0023] c0 is the conflict cost per unit time;

[0024] a j For binary decision variables, if the standby machine BM is activated... j Then a j The value is 1 if it is 1, otherwise it is 0.

[0025] x ij For binary decision variables, if workpiece J i Select in machine MM j If the processing is performed on the above, then mx ij The value is 1 if it is 1, otherwise it is 0.

[0026] y jki For binary decision variables, if That is, workpiece J k Compared to workpiece J i Process first, then y jki The value is 1 if it is 1, otherwise it is 0.

[0027] Preferably, based on the updated input parameters, a hybrid particle swarm optimization algorithm based on neighborhood search is used to completely reschedule the workpieces, obtain the global optimal solution, and decode the corresponding number of backup machines to be started, the production rescheduling scheme, the production cost of rescheduling, the cost of starting backup machines for rescheduling, and the cost of rescheduling conflicts; including:

[0028] S31. Set execution parameters, including at least: the maximum number of iterations Wl1 of the hybrid particle swarm algorithm based on neighborhood search, the population size Popsize of the algorithm, and the maximum number of iterations Wl2 of the local search.

[0029] S32. Generate an initial population X0 using a real number-based encoding method, generate an encoding of length l, assign workpieces to machines, use the weighted rescheduling total goal as the individual fitness value, and obtain and record the current optimal solution.

[0030] S33. Update the position of each individual particle based on the position update formula of the Particle Swarm Optimization (PSO) algorithm;

[0031] S34. Calculate the fitness value of each individual particle. The best position for each particle in the population is pbest, and the best position for the particle with the best scheduling target is gbest. Denote the best position gbest as π. Select the neighborhood structure NS based on roulette wheel selection. k And based on this neighborhood structure NS k Generate a neighborhood solution π′ of π.

[0032] S35. Based on a two-stage adaptive neighborhood search strategy, using the neighborhood structure NS k A local search is performed on the solution π′ to obtain the local optimum π″. The local optimum π″ is compared with the optimum π of the variable neighborhood search algorithm. If π″ is better than π, then π = π″. At the same time, the number of times the neighborhood structure successfully finds a local optimum is recorded, and the neighborhood structure NS is updated. k The selection probability is used to determine whether wl2 < Wl2 is true. If it is true, wl2 + 1 is assigned to wl2 and the process returns to step S34. Otherwise, step S36 is executed.

[0033] S36. Compare the optimal solution π of the variable neighborhood search algorithm with the worst individual in the initial population X0. If π is better than the worst individual in the initial population, replace the solution with π and compare π with the global optimal solution π. best If we compare, π is better than πbes t Then let π best =π;

[0034] S37. Determine if wl1 ≤ Wl1 is true. If yes, assign wl1 + 1 to wl1, use the updated population as the input for the next iteration, and return to step S33. Otherwise, output the global optimal solution π. best The corresponding number of standby machines to be started, production rescheduling scheme, rescheduling production cost, rescheduling standby machine start-up cost, and rescheduling conflict cost are decoded.

[0035] Preferably, the encoding rules of the neighborhood search-based hybrid particle swarm optimization algorithm include:

[0036] The LS code has a length of 1 and represents the assignment order of workpieces on the initial or standby machine. The real numbers in the LS code consist of an integer part and a fractional part, where the integer part... For the machine part number, decimal part The processing order of the workpieces is determined, with the workpieces having smaller decimal parts being processed earlier. The real number codes corresponding to different workpieces are not equal.

[0037] The decoding rules of the neighborhood search-based hybrid particle swarm optimization algorithm include:

[0038] S101. Given the processing sequence of the workpieces on each machine in the initial scheduling, This indicates that the initial schedule is for the c-th workpiece produced by the j-th machine. F represents the initial scheduling of the processing time for the c-th workpiece produced by the j-th machine. j For machine CM at time t j The set of already processed workpieces, Z j For the non-faulty machine CM at time t j The collection of workpieces being processed, S j Indicates the machine MM after rescheduling j The start time of the first workpiece processing on the machine, and the faulty machine CM j The estimated repair time is d;

[0039] S 102, Let j = 1;

[0040] S103, Let m = 1, for the processing machine CM j Determine the machine's CM j If a malfunction has occurred, proceed to step S104 if the machine is faulty, otherwise proceed to step S106.

[0041] S104, For processing machines CM j ,judge Is this valid? If so, update the machine's CM. j The above has completed the workpiece set F j If m+1 is assigned to m, continue to execute step S104; otherwise, execute step S105.

[0042] S105, For processing machines CM j ,judge Is this valid? If so, update the machine's CM. j The above has completed the workpiece set F j At this time, S j =t+d, execute step S108; otherwise, the workpiece being processed on the faulty machine will be reassigned. j =t+d, execute step S108;

[0043] S106, For processing machines CM j ,judge Is this valid? If so, update the machine's CM. j The above has completed the workpiece set F j If m+1 is assigned to m, continue executing S106; otherwise, execute step S107.

[0044] S107, For processing machines CM j ,judge Is this valid? If so, update the machine's CM. j The above has completed the workpiece set F j CM machine j The set of workpieces being processed is Z. j If S is an empty set, then S j =t, execute step S108; otherwise, output machine CM. j The set of workpieces being processed is Z. j ,at this time Execute step S108;

[0045] S108. When j≠u, assign j+1 to j and return to step S103. When j=u, execute step S109.

[0046] S109. Given the set F of all workpieces processed on all machines up to time t, and the set Z of all workpieces being processed on all non-faulty machines at time t, then set F = F1 ∪ ... ∪ F j ∪...∪F u j = [1, 2, ..., u], set Z = Z1 ∪ ... ∪ Z j ∪...∪Z u If j = [1, 2, ..., u] and j ≠ q, execute step S110;

[0047] S110. Generate a set of tasks to be scheduled J from the initial task set J, set F, and set Z. R , where set J R =J-(F∪Z);

[0048] S111, Order standby machine BM j Start time of machining of the first workpiece S j =t, execute step S112;

[0049] S112. Generate the three components of the overall rescheduling objective:

[0050] (1) Production cost f1 for rescheduling:

[0051]

[0052] Wherein, it represents the workpiece. In Machine MM j The unit time processing cost, P ij Indicates workpiece In Machine MM j On the processing time,

[0053]

[0054] (2) Energy consumption f2 for rescheduling standby machine startup:

[0055]

[0056] in

[0057] (3) Rescheduling deviation conflict cost f3:

[0058]

[0059] Indicates rescheduled workpiece Completion time, C i Indicates the initial scheduling of the workpiece. Completion time,

[0060]

[0061] y jki For decision variables, if δ k -[δ k 」<δ i -[δ i "That is, machine MM" j upper workpiece Compared to the workpiece Process first, then y jki The value is 1 if it is not 0 otherwise, C i The value is known.

[0062] Preferably, the process of obtaining the weighted rescheduling total objective as the individual fitness value in S32 is as follows:

[0063] Obtain the single-objective optimal solution f for each individual particle. 1max f 2max and f 3max ;

[0064] Weighted calculation of population fitness values:

[0065] f(x) = λ1f 1max +λ2f 2max +λ3f 3max

[0066] Where f(x) is the fitness function, λ1, λ2, and λ3 are the weights of production cost, rescheduling standby machine startup cost, and rescheduling deviation conflict cost, respectively, and λ1+λ2+λ3=1.

[0067] Preferably, step S33 updates the position of each individual particle based on the position update formula of the Particle Swarm Optimization (PSO) algorithm; including:

[0068]

[0069]

[0070] in, are the velocities of the m-th particle in the k-th and k+1-th iterations, respectively; The best position in the history of the m-th particle; gbest k The globally optimal position; ω represents the position of the m-th particle in the k-th and k+1-th iterations, respectively; c1 and c2 are acceleration constants; and r1 and r2 are random numbers.

[0071] Preferably, the neighborhood structure NS is used in S35. k A local search is performed on the solution π′ to obtain the locally optimal solution π″. This locally optimal solution π″ is then compared with the optimal solution π from the variable neighborhood search algorithm. If π″ is better than π, then π = π″. This includes:

[0072] S2501, Neighborhood Structure NS of Variable Neighborhood Search Algorithm k A two-stage adaptive neighborhood search method is used to optimize the local search process of the variable neighborhood search algorithm:

[0073] S2502. In the first stage, probability initialization is performed, and the number N of times each neighborhood structure successfully finds a local optimum is recorded. i Update the selection probability of each neighborhood according to the following formula: δ is a sufficiently small positive number;

[0074] S2503. In the second stage, the number N of successful neighborhood search updates for each individual, as counted in the first stage, will be used. i Reset to zero and use incremental learning to adjust the preset K. max The probabilistic model of the neighborhood structure is updated:

[0075]

[0076] Where SPi(s) and SPi(s+1) are the probabilities of the i-th neighborhood structure being selected in the s-th and s+1-th neighborhood searches, respectively.

[0077] A high-end equipment production rescheduling optimization system considering machine failure scenarios includes:

[0078] The setting module is used to set input parameters based on workpiece data, machine data and rescheduling data during the production process of high-end equipment workshop. These parameters include at least the workpiece set, machine set, workpiece processing start and end time, standby machine start-up cost, conflict cost, machine failure time point and fixed machine maintenance duration.

[0079] The update module is used to update the workpiece set, machine set, and workpiece processing start time after a machine failure occurs.

[0080] The solution module is used to perform a complete rescheduling of the workpiece based on the updated input parameters and a hybrid particle swarm optimization algorithm based on neighborhood search, to obtain the global optimal solution, and decode the corresponding number of standby machines to start, production rescheduling scheme, production cost of rescheduling, cost of starting standby machines for rescheduling, and cost of rescheduling conflicts.

[0081] A storage medium storing a computer program for optimizing the rescheduling of high-end equipment production under machine failure conditions, wherein the computer program causes a computer to execute the optimization method for optimizing the rescheduling of high-end equipment production under machine failure conditions as described above.

[0082] An electronic device, comprising:

[0083] One or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including methods for performing the high-end equipment production rescheduling optimization method considering machine failure scenarios as described above.

[0084] (III) Beneficial Effects

[0085] This invention provides a method, system, storage medium, and electronic device for optimizing the rescheduling of high-end equipment production under machine failure conditions. Compared with the prior art, it has the following advantages:

[0086] This invention first sets input parameters based on workpiece data, machine data, and rescheduling data from the production process in a high-end equipment workshop. Then, after a machine failure, the workpiece set, machine set, and workpiece processing start time are updated. Finally, based on the updated input parameters, a hybrid particle swarm optimization algorithm based on neighborhood search is used to completely reschedule the workpieces, obtaining the global optimal solution. The corresponding number of backup machines to be started, the production rescheduling scheme, the production cost of rescheduling, the startup cost of backup machines, and the rescheduling conflict cost are decoded. This invention considers the multi-machine collaborative rescheduling problem under machine failure conditions in the workshop and comprehensively considers the optimization objectives of rescheduling processing time and cost. A workpiece sequential scheduling model for a flexible workshop with manufacturing interruptions is constructed, and an improved particle swarm optimization algorithm is proposed to solve this problem. This effectively reduces production costs and enhances the enterprise's ability to cope with sudden production interruptions. Attached Figure Description

[0087] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0088] Figure 1 A block diagram illustrating a method for optimizing the production rescheduling of high-end equipment under machine failure conditions, as provided in an embodiment of the present invention.

[0089] Figure 2 A flowchart of a hybrid particle swarm optimization algorithm based on neighborhood search provided in an embodiment of the present invention;

[0090] Figure 3 This is an example of encoding provided for an embodiment of the present invention. Detailed Implementation

[0091] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are described clearly and completely. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0092] This application provides a method, system, storage medium, and electronic device for optimizing the rescheduling of high-end equipment production under machine failure conditions, thereby solving the technical problem of workshop rescheduling with backup machines.

[0093] The technical solution in this application is to solve the above-mentioned technical problems, and the general idea is as follows:

[0094] This invention first acquires task data during the rescheduling process of high-end equipment manufacturing, and sets input parameters for a hybrid particle swarm optimization (PSO) algorithm based on neighborhood search based on the task data during the rescheduling process of high-end equipment supply manufacturing; then, it sets the execution parameters of the algorithm and initializes the initial solution of the algorithm; next, it updates the position of each particle based on the position update formula of the PSO algorithm; it calculates the fitness value of each particle, selects the neighborhood structure to generate the neighborhood solution of the optimal particle in the swarm; it generates the neighborhood solution based on the neighborhood structure and performs a local search operation.

[0095] This invention reschedules the production process of multiple machines interrupted in the high-end equipment manufacturing industry, shortening the production time of enterprises while reducing production energy consumption, thereby effectively improving the production scheduling and management level of high-end equipment enterprises and their ability to cope with sudden production interruptions.

[0096] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0097] Example 1:

[0098] like Figure 1 As shown, this embodiment of the invention provides a method for optimizing the rescheduling of high-end equipment production under machine failure conditions, including:

[0099] S1. Based on workpiece data, machine data and rescheduling data during the production process in the high-end equipment workshop, set input parameters, including at least the workpiece set, machine set, workpiece processing start and end time, standby machine start-up cost, conflict cost, machine failure time point and fixed machine maintenance time.

[0100] S2. After a machine failure occurs, update the workpiece set, machine set, and workpiece processing start time;

[0101] S3. Based on the updated input parameters, a hybrid particle swarm optimization algorithm based on neighborhood search is used to completely reschedule the workpieces, obtain the global optimal solution, and decode the corresponding number of standby machines to start, production rescheduling scheme, production cost of rescheduling, cost of starting standby machines for rescheduling, and cost of rescheduling conflicts.

[0102] This invention addresses the multi-machine collaborative rescheduling problem under machine failure conditions in the workshop. It comprehensively considers the optimization objectives of rescheduling processing time and cost, constructs a workpiece sequential scheduling model for a flexible workshop during manufacturing interruptions, and proposes an improved particle swarm optimization algorithm to solve the problem. This approach can effectively reduce manufacturing costs and improve the enterprise's ability to cope with sudden production interruptions.

[0103] The following will detail each step of the above solution:

[0104] In step S1, based on workpiece data, machine data, and rescheduling data during the production process in the high-end equipment workshop, input parameters are set, including at least the workpiece set, machine set, workpiece processing start and end times, standby machine startup cost, conflict cost, machine failure time point, and fixed machine maintenance duration.

[0105] This step involves obtaining workpiece data, machine data, and rescheduling data before the interruption of high-end equipment production through manual input and other methods. Based on the obtained data, the input parameters of the hybrid algorithm of particle swarm optimization and variable neighborhood search are set.

[0106] Specifically, the input parameters include:

[0107] The set of n initial workpieces, J = [J1, J2, ..., Jn], consists of a workpiece set J = [J1, J2, ..., Jn]. n ];

[0108] A set of workpieces consisting of l unprocessed workpieces

[0109] The set of machines MM = {MM1, MM2, ..., MM} u BM u+1 BM u+2 , ..., BM p}, where MM j For the machines that perform the initial production operation, u is the number of machines for the initial production operation, BM j p represents the number of newly activated standby machines to be rescheduled.

[0110] Assume CM j CM (China Machine Tool) q A fault occurs, q∈[1,2,...,u];

[0111] CM q The time interval t from the start of processing to the occurrence of failure follows an exponential distribution, and the failed machine CM q The maintenance time d follows an exponential distribution;

[0112] Machine CM up to time t j The set of processed workpieces F j The set of all workpieces processed on all machines up to time t;

[0113] Non-faulty machine CM at time t j The set of workpieces being processed is Z. j Z is the set of all workpieces being processed on all non-faulty machines at time t;

[0114] P ij For workpiece J i In Machine MM j The time spent processing on top, C i Indicates the initial scheduling of workpiece J i Processing completion time Indicates rescheduling of workpiece J i Processing completion time; S j For rescheduled machine MM j Start time of the first workpiece processing;

[0115] c ij For workpiece J i In Machine MM j The unit time processing cost;

[0116] ε BM Startup cost for each backup machine;

[0117] c0 is the conflict cost per unit time;

[0118] a j For binary decision variables, if the standby machine BM is activated... j Then a j The value is 1 if it is not 1, otherwise it is 0.

[0119] x ij For binary decision variables, if workpiece J i Select in machine MM j If the processing is performed on the above, then mx ij The value is 1 if it is not 1, otherwise it is 0.

[0120] y jki For binary decision variables, if That is, workpiece J k Compared to workpiece J i Process first, then y jki The value is 1 if it is 1, otherwise it is 0.

[0121] In step S2, after a machine failure occurs, the workpiece set, machine set, and workpiece processing start time are updated.

[0122] It should be noted that, in this embodiment of the invention, after a machine failure occurs, processing resumes once the failed machine has been repaired. Workpieces being processed on the failed machine are reassigned, while workpieces being processed on non-faulty machines continue processing. The workpieces are completely rescheduled, and the workpiece set, machine set, and workpiece processing start time are updated accordingly.

[0123] In step S3, based on the updated input parameters, a hybrid particle swarm optimization algorithm based on neighborhood search is used to completely reschedule the workpieces, obtain the global optimal solution, and decode the corresponding number of standby machines to start, the production rescheduling scheme, the production cost of rescheduling, the start cost of rescheduling standby machines, and the cost of rescheduling conflicts.

[0124] The present invention proposes an improved particle swarm optimization algorithm, namely a hybrid particle swarm optimization algorithm based on neighborhood search, to solve the multi-machine production collaborative rescheduling problem under the condition of machine failure in the workshop, which can effectively reduce production and manufacturing costs.

[0125] like Figure 2 As shown, Figure 2 A flowchart of a hybrid particle swarm optimization algorithm based on neighborhood search is presented. Correspondingly, combined with... Figure 2 This step specifically includes:

[0126] S31. Set execution parameters, including at least: the maximum number of iterations Wl1 of the hybrid particle swarm algorithm based on neighborhood search, the population size Popsize of the algorithm, and the maximum number of iterations Wl2 of the local search.

[0127] S32. Generate an initial population X0 using a real number-based encoding method, generate an encoding of length l, assign workpieces to machines, use the weighted rescheduling total goal as the individual fitness value, and obtain and record the current optimal solution.

[0128] S33. Update the position of each individual particle based on the position update formula of the Particle Swarm Optimization (PSO) algorithm;

[0129] S34. Calculate the fitness value of each individual particle. The best position for each particle in the population is pbest, and the best position for the particle with the best scheduling target is gbest. Denote the best position gbest as π. Select the neighborhood structure NS based on roulette wheel selection. k And based on this neighborhood structure NS k Generate a neighborhood solution π′ of π;

[0130] S35. Based on a two-stage adaptive neighborhood search strategy, using the neighborhood structure NS k A local search is performed on the solution π′ to obtain the local optimum π″. The local optimum π″ is compared with the optimum π of the variable neighborhood search algorithm. If π″ is better than π, then π = π″. At the same time, the number of times the neighborhood structure successfully finds a local optimum is recorded, and the neighborhood structure NS is updated. k The selection probability is used to determine whether wl2 < Wl2 is true. If it is true, wl2 + 1 is assigned to wl2 and the process returns to step S34. Otherwise, step S36 is executed.

[0131] S36. Compare the optimal solution π of the variable neighborhood search algorithm with the worst individual in the initial population X0. If π is better than the worst individual in the initial population, replace the solution with π and compare π with the global optimal solution π. best If we compare, π is better than π best Then let π best =π;

[0132] S37. Determine if wl1 ≤ Wl1 is true. If yes, assign wl1 + 1 to wl1, use the updated population as the input for the next iteration, and return to step S33. Otherwise, output the global optimal solution π. best The corresponding number of standby machines to be started, production rescheduling scheme, rescheduling production cost, rescheduling standby machine start-up cost, and rescheduling conflict cost are decoded.

[0133] In an optional implementation, the encoding rules of the hybrid particle swarm optimization algorithm based on neighborhood search provided in this embodiment of the invention specifically include:

[0134] The LS code has a length of 1 and represents the assignment order of workpieces on the initial or standby machine. The real numbers in the LS code consist of an integer part and a fractional part, where the integer part... For the machine part number, decimal part The processing order of the workpieces is determined, with the workpieces having smaller decimal parts being processed earlier. The real number codes corresponding to different workpieces are not equal.

[0135] Furthermore, the completed workpieces and those being processed by non-faulty machines are removed from the workpiece set and their quantity is recorded as nl. A code of length l is generated, and the workpieces are assigned to the initial working machine or the standby machine.

[0136] like Figure 3 As shown, Figure 3 An encoding example is provided.

[0137] Correspondingly, the decoding process of the hybrid particle swarm algorithm based on neighborhood search provided in this embodiment of the invention specifically includes:

[0138] S101. Given the processing sequence of the workpieces on each machine in the initial scheduling, This indicates that the initial schedule is for the c-th workpiece produced by the j-th machine. F represents the initial scheduling of the processing time for the c-th workpiece produced by the j-th machine. j For machine CM at time t j The set of already processed workpieces, Z j For the non-faulty machine CM at time t j The collection of workpieces being processed, S j Indicates the machine MM after rescheduling j The start time of the first workpiece processing on the machine, and the faulty machine CM j The estimated repair time is d;

[0139] S102, Let j = 1;

[0140] S103, Let m = 1, for the processing machine CM j Determine the machine's CM j If a malfunction has occurred, proceed to step S104 if the machine is faulty, otherwise proceed to step S106.

[0141] S104, For processing machines CM j ,judge Is this valid? If so, update the machine's CM. j The above has completed the workpiece set F j If m+1 is assigned to m, continue to execute step S104; otherwise, execute step S105.

[0142] S105, For processing machines CM j ,judge Is this valid? If so, update the machine's CM. j The above has completed the workpiece set F j At this time, S j =t+d, execute step S108; otherwise, the workpiece being processed on the faulty machine will be reassigned. j =t+d, execute step S108;

[0143] S106, For processing machines CM j ,judge Is this valid? If so, update the machine's CM. j The above has completed the workpiece set F j If m+1 is assigned to m, continue executing S106; otherwise, execute step S107.

[0144] S107, For processing machines CM j ,judge Is this valid? If so, update the machine's CM. j The above has completed the workpiece set F j CM machine j The set of workpieces being processed is Z. j If S is an empty set, then S j =t, execute step S108; otherwise, output machine CM. j The set of workpieces being processed is Z. j ,at this time Execute step S108;

[0145] S108. When j≠u, assign j+1 to j and return to step S103. When j=u, execute step S109.

[0146] S109. Given the set F of all workpieces processed on all machines up to time t, and the set Z of all workpieces being processed on all non-faulty machines at time t, then set F = F1 ∪ ... ∪ F j ∪...∪F u j = [1, 2, ..., u], set Z = Z1 ∪ ... ∪ Z j ∪...∪Z u If j = [1, 2, ..., u] and j ≠ q, execute step S110;

[0147] S110. Generate a set of tasks to be scheduled J from the initial task set J, set F, and set Z. R , where set J R =J-(F∪Z);

[0148] S111, Order standby machine BMj Start time of machining of the first workpiece S j =t, execute step S112;

[0149] S112. Generate the three components of the overall rescheduling objective:

[0150] (1) Production cost f1 for rescheduling:

[0151]

[0152] Wherein, it represents the workpiece. In Machine MM j The unit time processing cost, P ij Indicates workpiece In Machine MM j On the processing time,

[0153]

[0154] (2) Energy consumption f2 for rescheduling standby machine startup:

[0155]

[0156] in

[0157] (3) Rescheduling deviation conflict cost f3:

[0158]

[0159] Indicates rescheduled workpiece Completion time, C i Indicates the initial scheduling of the workpiece. Completion time,

[0160]

[0161] y jki For decision variables, if That is, machine MM j upper workpiece Compared to the workpiece Process first, then y jki The value is 1 if it is not 0 otherwise, C i The value is known.

[0162] It should be noted that the cost of deviation conflict after rescheduling is an important indicator for measuring the robustness of rescheduling. A higher value means that more processes are moved, which makes it more inconvenient to reschedule production of workpieces.

[0163] Based on the aforementioned encoding and decoding rules, in an optional implementation, the process of obtaining the weighted rescheduling target as the individual fitness value in S32 is as follows:

[0164] Obtain the single-objective optimal solution f for each individual particle. 1max f 2max and f 3max ;

[0165] Weighted calculation of population fitness values:

[0166] f(x) = λ1f 1max +λ2f 2max ++λ3f 3max

[0167] Where f(x) is the fitness function, λ1, λ2, and λ3 are the weights of production cost, rescheduling standby machine startup cost, and rescheduling deviation conflict cost, respectively, and λ1+λ2+λ3=1.

[0168] In an optional implementation, step S33 updates the position of each individual particle based on the position update formula of the Particle Swarm Optimization (PSO) algorithm; including:

[0169]

[0170]

[0171] in, are the velocities of the m-th particle in the k-th and k+1-th iterations, respectively; The best position in the history of the m-th particle; gbest k The globally optimal position; ω represents the position of the m-th particle in the k-th and k+1-th iterations, respectively; c1 and c2 are acceleration constants; and r1 and r2 are random numbers.

[0172] In this embodiment of the invention, the feasible solution π is in the selected neighborhood structure NS k A transformation is performed to jump the initial solution to another solution in the feasible region, avoiding repeated iterations and preventing the algorithm from getting trapped in local optima. This takes into account the complexity of the high-end equipment production rescheduling process and the requirements of the constraints.

[0173] For example, the following four neighborhood structures are designed:

[0174] Neighborhood Structure 1: Randomly generate the real number encoding value of the workpiece with the largest deviation in processing completion time before and after rescheduling.

[0175] Neighborhood structure 2: Define a variable a, randomly generate an integer in the interval [1, l], and assign it to a. Assign the workpiece at position a in the sequence LS to the machine with the shortest processing time.

[0176] Neighborhood structure 3: Define variables a and b, randomly generate two integers in the interval [1, l], assign them to a and b, where a < b. Swap the encodings at positions a and b in the sequence LS.

[0177] Neighborhood structure 4: Define variables a and b, randomly generate two integers in the interval [1, l], assign them to a and b, where a < b, and reverse the encoding between positions a and b in the sequence LS.

[0178] In addition, the embodiment of the present invention introduces an adaptive incremental learning process in the variable neighborhood search algorithm and continuously updates it during the iteration process. Specifically:

[0179] In an optional embodiment, the neighborhood structure NS is used in S35 k to perform local search on the solution π′ to obtain the local optimal solution π″, and compare the local optimal solution π″ with the optimal solution π of the variable neighborhood search algorithm. If π″ is better than π, then let π = π″; including:

[0180] S2501. The neighborhood structure NS of the variable neighborhood search algorithm k , and is optimized by adopting a two-stage adaptive neighborhood search method during the local search process of the variable neighborhood search algorithm:

[0181] S2502. In the first stage, perform probability initialization, and record the number of times N that each neighborhood structure successfully searches to the local optimal solution i , and update the selection probability of each neighborhood according to the following formula: δ is a sufficiently small positive number;

[0182] S2503. In the second stage, clear the number of times N that each neighborhood search successfully updates the individual statistically in the first stage i , and update the probability model of the preset K max types (which can be the 4 types listed above) of neighborhood structures in an incremental learning manner:

[0183]

[0184] where Spi(s) and Spi(s + 1) are the probabilities that the i-th neighborhood structure is selected in the s-th and s + 1-th neighborhood searches respectively.

[0185] So far, the embodiment of the present invention has completed all the processes of the high-end equipment production rescheduling optimization method considering the machine failure situation.

[0186] Example 2:

[0187] This invention provides a high-end equipment production rescheduling optimization system considering machine failure scenarios, comprising:

[0188] The setting module is used to set input parameters based on workpiece data, machine data and rescheduling data during the production process of high-end equipment workshop. These parameters include at least the workpiece set, machine set, workpiece processing start and end time, standby machine start-up cost, conflict cost, machine failure time point and fixed machine maintenance duration.

[0189] The update module is used to update the workpiece set, machine set, and workpiece processing start time after a machine failure occurs.

[0190] The solution module is used to perform a complete rescheduling of the workpiece based on the updated input parameters and a hybrid particle swarm optimization algorithm based on neighborhood search, to obtain the global optimal solution, and decode the corresponding number of standby machines to start, production rescheduling scheme, production cost of rescheduling, cost of starting standby machines for rescheduling, and cost of rescheduling conflicts.

[0191] Example 3:

[0192] This invention provides a storage medium storing a computer program for optimizing the rescheduling of high-end equipment production under machine failure conditions. The computer program causes a computer to execute the high-end equipment production rescheduling optimization method under machine failure conditions as described in Embodiment 1.

[0193] Example 4:

[0194] This invention provides an electronic device, comprising:

[0195] One or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including methods for performing the high-end equipment production rescheduling optimization method considering machine failure scenarios as described in Example 1.

[0196] It is understood that the high-end equipment production rescheduling optimization system, storage medium and electronic device considering machine failure scenarios provided in the embodiments of the present invention correspond to the high-end equipment production rescheduling optimization method considering machine failure scenarios provided in the embodiments of the present invention. The explanation, examples and beneficial effects of the relevant contents can be referred to the corresponding parts of the high-end equipment production rescheduling optimization method considering machine failure scenarios, and will not be repeated here.

[0197] In summary, compared with existing technologies, it has the following beneficial effects:

[0198] This invention addresses the multi-machine collaborative rescheduling problem under machine failure conditions in the workshop. It comprehensively considers the optimization objectives of rescheduling processing time and cost, constructs a workpiece sequential scheduling model for a flexible workshop during manufacturing interruptions, and proposes an improved particle swarm optimization algorithm to solve the problem. This approach can effectively reduce manufacturing costs and improve the enterprise's ability to cope with sudden production interruptions.

[0199] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0200] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for high-end equipment production rescheduling optimization considering machine failure scenarios, characterized in that, The method comprises the following steps: Based on the workpiece data, machine data and rescheduling data in the production process of the high-end equipment workshop, input parameters are set, including at least a workpiece set, a machine set, workpiece processing start and end times, standby machine start-up costs, conflict costs, machine failure time points and machine maintenance fixed time lengths; After a machine failure occurs, the workpiece set, machine set and workpiece processing start time are updated; Based on the updated input parameters, a hybrid particle swarm algorithm based on neighborhood search is used to completely reschedule the workpieces, a global optimal solution is obtained, and the corresponding number of standby machines started, production rescheduling scheme, rescheduling production cost, rescheduling standby machine start-up cost and rescheduling conflict cost are decoded; The method comprises the following steps: S31, setting execution parameters, including at least the maximum number of iterations Wl1 of the hybrid particle swarm algorithm based on neighborhood search, the population size Popsize of the algorithm and the maximum number of iterations Wl2 of local search; S32, generating an initial population X0 using a real number-based encoding method, generating an encoding with a length of l, assigning workpieces to machines, and obtaining and recording the current optimal solution by taking the weighted rescheduling total target as the individual fitness value; The input parameters specifically include: S34, calculate the fitness value of each particle individual, the optimal pbest of each individual in the population is the current position of the individual, the particle individual best for scheduling target in the population is the group optimal position gbest, record the group optimal particle individual abest as π, select the neighborhood structure NS based on the roulette method k , and generate a neighborhood solution π' of π based on the neighborhood structure NS k ​ S35, based on two-stage adaptive neighborhood search strategy, using neighborhood structure NS k A local search is performed on the solution π' to obtain a local optimal solution π", and the local optimal solution π" is compared with the optimal solution π of the variable neighborhood search algorithm. If π" is better than π, then π = π", and the number of successful times of searching for the local optimal solution by the neighborhood structure is recorded, and the neighborhood structure NS is updated k The selection probability of the neighborhood structure NS is updated, and it is determined whether wl2 < wl2 holds true. If yes, wl2 + 1 is assigned to wl2, and the step S34 is returned. Otherwise, the step S36 is executed. S36, compare the optimal solution of the variable neighborhood search algorithm with the worst individual in the initial population X0, if π is better than the worst individual in the initial population, replace the solution with π, and compare π with the global optimal solution π best , if π is better than π best , let π best = π; S37, judging whether wl1≤wl1 is true, if yes, assigning wl1+1 to wl1, taking the updated population as the input of next iteration, returning to step S33, otherwise, outputting the global optimal solution π best and the corresponding number of starting backup machines, production rescheduling scheme, rescheduling production cost, rescheduling backup machine starting cost and rescheduling conflict cost are obtained by decoding.

2. The method for high-end equipment production rescheduling optimization considering machine failure scenarios according to claim 1, characterized in that, c0 is the conflict cost per unit time; A workpiece set J = [J1, J2,..., Jn] consisting of n initial workpieces n ] a workpiece set consisting of l workpieces Machine set MM = {MM1, MM2,..., MM u ,BM u+1 ,BM u+2 ,…,BM p} where MM j is a machine that performs an initial production job, u is the number of machines that perform the initial production job, BM j is a newly started backup machine that is rescheduled, and p is the number of newly started backup machines that are rescheduled. Assume CM j Machine CM q Fails, q e [1,2,…,u] CM q The time interval t from the start of processing to the occurrence of a fault obeys an exponential distribution, the repair duration d of the faulty machine CM q obeys an exponential distribution; Machine CM up to time t j The set of processed workpieces F j The set of all workpieces processed on all machines up to time t; t time non-faulty machine CM j upper being processed workpiece set Z j t time all non-faulty machine being processed workpiece set Z P ij for the workpiece J i the time spent on the machine MM j for processing, C i indicates the initial scheduling workpiece J i processing completion time, indicates the rescheduling workpiece J i processing completion time; S j for the first workpiece processing start time on the rescheduled machine MM j for the first workpiece processing start time on the rescheduled machine MM c ij for the workpiece J i on the machine MM j per time processing cost; ε BM cost to start each spare machine; 3. The high-end equipment production rescheduling optimization method considering machine failure, according to claim 1, wherein a j For binary decision variables, if the backup machine BM is enabled j then a j value is 1, otherwise 0; x ij For binary decision variables, if workpiece J i Select in machine MM j If the processing is performed on the above, then mx ij The value is 1 if it is not 1, otherwise it is 0. y jki is a binary decision variable, if i.e. the job J k is processed before the job J i , then y jki has the value 1, otherwise 0. The encoding rules of the hybrid particle swarm algorithm based on neighborhood search include: The decoding process is as follows: The coding LS has a length of l, and the coding LS has a meaning of assignment sequence of the workpieces on the initial working machine or the standby machine; the real number in the coding LS contains an integer part and a decimal part, wherein the integer part is a number of the working machine, and the decimal part determines the processing sequence of the workpieces, and the smaller the decimal part of the workpiece is, the earlier the processing position of the workpiece is, wherein the real number coding values corresponding to different workpieces are not equal. S2202, the decoding process is to calculate the rescheduling scheme, the set of standby machines enabled, the rescheduling production cost and the rescheduling standby machine start-up energy consumption according to the actual situation; including the following operations: S102, let j = 1; S101, the processing sequence of the workpieces on each machine in the initial scheduling is known, represents the cth workpiece produced by the jth machine in the initial scheduling, represents the processing time of the cth workpiece produced by the jth machine in the initial scheduling, F j is the set of workpieces processed on the machine CM j before t, Z j is the set of workpieces processed on the machine CM j before t, S j represents the first workpiece processed on the machine MM j after the rescheduling, the start time of the first workpiece processed on the machine CM j is the predicted maintenance duration, d; S108, when j ≠ u, assign j + 1 to j, return to step S103, and when j = u, execute step S109; S103, let m = 1, for the machining machine CM j , judge whether the machine CM j is in failure, if it is a failure machine, execute step S104, otherwise, execute step S106; S104, for the machining machine CM j , judging whether the following is true, and if so, updating the set of workpieces F j on the machine CM j , assigning m+1 to m, and continuing with step S104, otherwise executing step S105; S105, for the machining machine CM j , it is judged if it is true, the machine CM j on which the set of workpieces F j has been completed is updated, at this time S j = t + d, step S108 is executed, otherwise, the workpiece being processed on the faulty machine will be reassigned, at this time S j = t + d, step S108 is executed; S106, for the machining machine CM j , it is judged whether it is true, if it is true, the set F j of workpieces completed on the machine CM j is updated, m+1 is assigned to m, and the step S106 is continued, otherwise the step S107 is executed; S107, For processing machines CM j ,judge Is this valid? If so, update the machine's CM. j The above has completed the workpiece set F j CM machine j The set of workpieces being processed is Z. j If S is an empty set, then S j =t, execute step S108; otherwise, output machine CM. j The set of workpieces being processed is Z. j ,at this time Execute step S108; S112, generate three components of the rescheduling total target respectively: S109, the set F of all the processed workpieces on all the machines at time t, the set Z of all the workpieces being processed on all the non-faulty machines at time t, then the set F=F1∪…∪Z j ∪…∪F u , j=[1, 2, …, u], the set Z=Z1∪…∪F j ∪…∪Z u , j=[1, 2, …, u] and j≠q, step S110 is executed; S110, generating a set J of to-be-scheduled jobs from the initial set J of jobs, the set F, and the set Z R , wherein the set J R = J - (F U Z); S111, the standby machine BM j the first workpiece processing start time S j = t, execute step S112; (1) Rescheduling production cost f1: (2) Rescheduling standby machine start-up energy consumption f2: Wherein, it represents the workpiece. In Machine MM j The unit time processing cost, P ij Indicates workpiece In Machine MM j On the processing time, (3) Rescheduling deviation conflict cost f3: wherein The process of obtaining the weighted rescheduling total target as the individual fitness value in S32 is as follows: re-scheduled job completion time, C i initially scheduled job completion time, y jki is a decision variable, if i.e. the machine MM j is above the workpiece is below the workpiece is processed first, y jki has the value 1, otherwise 0, C i is known.

4. The method of high-end equipment production rescheduling optimization considering machine failure scenarios of claim 3, wherein, Weighted population fitness value calculation: obtaining each target optimal solution f of each particle individual 1max , f 2max , and f 3max ; Wherein, f(x) is the fitness function, λ1, λ2, λ3 are the weights of the rescheduling production cost, the rescheduling standby machine start-up cost and the rescheduling deviation conflict cost respectively, and λ1+λ2+λ3=1. f(x) = λ1f 1max + λ2f 2max + λ3f 3max The position update formula of the particle swarm optimization algorithm PSO in S33 updates the position of each particle individual; including:

5. The method for high-end equipment production rescheduling optimization considering machine failure scenarios of claim 1, wherein, A device for executing the high-end equipment production rescheduling optimization method considering machine failure, according to claim 1, comprising: wherein, vm(k) and vm(k+1) are the velocity of the mth particle individual at the kth and k+1th iteration, respectively; gbestis the global best position; k gbestis the global best position; xm(k) and xm(k+1) are the position of the mth particle individual at the kth and k+1th iteration, respectively; ω denotes the inertia coefficient; c1and c2are the acceleration constants; and r1and r2are random numbers.

6. The method for high-end equipment production rescheduling optimization considering machine failure scenarios of claim 1, wherein, The S35 uses a neighborhood structure NS k A local search is performed on the solution π' to obtain a local optimal solution π", and the local optimal solution π" is compared with the optimal solution π of the variable neighborhood search algorithm, and if π" is better than π, then π = π"; comprising: S2501, the neighborhood structure NS of the variable neighborhood search algorithm k In the local search process of the variable neighborhood search algorithm, a two-stage adaptive neighborhood search method is used for optimization: S2502, the first stage carries out probability initialization, and records the number N of times that each neighborhood structure successfully searches for a local optimal solution i The selection probability of each neighborhood is updated according to the following formula: δ is a sufficiently small positive number; S2503、In the second stage, the number of times N of updating the individual by each neighborhood search success in the first stage is counted i Clearing and using the incremental learning method to update the preset K max The probability model of the neighborhood structure is updated: where SP i (s), SP i (s+1) are the probabilities of the i-th neighborhood structure being selected in the s, s+1 neighborhood search, respectively.

7. A high-end equipment production rescheduling optimization system considering machine failure scenarios, characterized in that, ​ The setting module is configured to set input parameters based on workpiece data, machine data, and rescheduling data in a high-end equipment workshop production process, the input parameters at least including a workpiece set, a machine set, workpiece processing start and end times, a standby machine startup cost, a conflict cost, a machine failure time point, and a machine maintenance fixed duration; The updating module is configured to update the workpiece set, the machine set, and the workpiece processing start time after a machine failure occurs; The solving module is configured to perform complete rescheduling on the workpieces based on the updated input parameters by using a hybrid particle swarm algorithm based on neighborhood search, to obtain a global optimal solution, and to decode to obtain corresponding numbers of standby machines to start, a production rescheduling scheme, a rescheduling production cost, a rescheduling standby machine startup cost, and a rescheduling conflict cost.

8. A storage medium, characterized by The computer program for high-end equipment production rescheduling optimization considering a machine failure scenario is stored in the computer readable storage medium, and the computer program causes the computer to perform the high-end equipment production rescheduling optimization method considering the machine failure scenario.

9. An electronic device, comprising: The computer readable storage medium comprises: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs comprise a program for performing the high-end equipment production rescheduling optimization method considering the machine failure scenario.

Images