Multi-objective flexible workshop dynamic scheduling method, device and electronic equipment under dual resource constraints

Abstract

The present invention discloses a multi-objective flexible workshop dynamic scheduling method, device and electronic equipment under dual resource constraints. First, the dual resource constraints and production assumptions of the production workshop are determined, and a multi-objective static scheduling model of the flexible workshop under dual resource constraints is established; then the processing information of the orders to be processed and the worker scheduling information are obtained from the workshop, and a pre-scheduling plan is generated based on a hybrid particle swarm algorithm for workshop production execution; a dynamic event trigger mechanism is used to monitor the real-time production status of the workshop to determine whether a disturbance is triggered; for two dynamic events, machine failure and new workpiece arrival, dynamic event parameters are determined; finally, a hybrid particle swarm algorithm is used to reschedule the processing tasks, and the optimal scheduling decision plan, completion cycle and workshop energy consumption under dual resource constraints are obtained. The present invention can effectively reduce the impact of interference on workshop stability and improve production robustness and workshop efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of flexible job shop scheduling, and in particular relates to a multi-objective flexible job shop dynamic scheduling method, device and electronic equipment under dual resource constraints. Background Art

[0002] Flexible workshop scheduling technology is designed to solve the problem of workshop resource allocation. In-depth research on workshop scheduling technology has great engineering application value. From classic workshops to flexible workshops, from single-resource workshop scheduling to dual-resource workshop scheduling, researchers at home and abroad have been committed to expanding workshop scheduling models to meet production requirements. The dual-resource workshop scheduling model considers both machine resources and human resources on the basis of the classic workshop scheduling model, which is more in line with production reality and is a hot topic in the field of workshop scheduling research today. Processing workpieces under dual resource constraints has high requirements for workers and machine resources, the production constraint relationship is obvious, and the solution complexity is high.

[0003] The Chinese patent application "A method, system, equipment and medium for scheduling large and complex product workshops" (202310810708.2) uses a multi-objective genetic algorithm to optimize the solution of large and complex product assembly and debugging workshop scheduling problems, obtains scheduling decisions and completion times, and effectively improves the solution accuracy. However, this method has the following shortcomings: 1. Worker constraints are not considered during model establishment. 2. The optimization algorithm is relatively simple and the solution efficiency is low. At present, the flexible workshop scheduling problem under dual resource constraints mainly has two major problems: difficulty in model establishment and frequent dynamic events. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the prior art and provide a multi-objective flexible workshop dynamic scheduling method, device and electronic equipment under dual resource constraints. First, the dual resource constraints and production assumptions of the production workshop are determined, and a multi-objective static scheduling model of the flexible workshop under dual resource constraints is established; then the processing information of the orders to be processed and the worker schedule information are obtained from the workshop, and a pre-scheduling plan is generated based on an improved hybrid particle swarm algorithm for workshop production execution; a dynamic event trigger mechanism is used to monitor the real-time production status of the workshop to determine whether a disturbance is triggered; for two dynamic events, machine failure and new workpiece arrival, dynamic event parameters are determined; finally, the improved hybrid particle swarm algorithm is used to reschedule the processing tasks, and the optimal scheduling decision plan, completion cycle and workshop energy consumption under dual resource constraints are obtained, thereby effectively solving the problems of low solution efficiency and dynamic interference in the prior art.

[0005] To achieve the above objectives, the present invention adopts the following technical solutions.

[0006] A multi-objective flexible workshop dynamic scheduling method under dual resource constraints of the present invention comprises the following steps:

[0007] Step 1: Determine the dual resource constraints of machines and workers in the production workshop and the production assumptions; determine the multi-objective optimization indicators, including the maximum completion time and workshop energy consumption, and obtain the objective function; analyze the multi-objective dynamic scheduling characteristics of the dual-resource flexible workshop, and establish a multi-objective static scheduling model for the flexible workshop under dual resource constraints;

[0008] Step 2: Obtain the processing information of the orders to be processed and the worker scheduling information from the workshop, and generate a pre-scheduling plan based on the improved hybrid particle swarm algorithm for the workshop production execution. The specific process includes: obtaining the input information of the flexible workshop multi-objective static scheduling model from the workshop, including the processing machines, processing time, processing power and worker information of different workpieces that can operate the processing machines; determining the multi-objective optimization indicators, including the maximum completion time and workshop energy consumption, to obtain the objective function; using the collaborative optimization strategy to improve the hybrid particle swarm algorithm to obtain the improved hybrid particle swarm algorithm;

[0009] Step 3: Use a dynamic event trigger mechanism to monitor the real-time production status of the workshop and determine whether a disturbance is triggered;

[0010] Step 4: Determine dynamic event parameters for two dynamic events: machine failure and new workpiece arrival;

[0011] Step 5: Use the improved hybrid particle swarm algorithm to reschedule the processing tasks and obtain the optimal scheduling decision plan, completion cycle and workshop energy consumption under dual resource constraints;

[0012] The dual resource constraints of production workshop machines and workers and production assumptions include:

[0013] Assumptions:

[0014] The multi-objective dual-resource flexible workshop scheduling problem is described as follows: w workers operate m machines to complete the processing task of n workpieces, each workpiece consists of at least one process, any process can be processed on one or more machines, and each worker can operate one or more machines; different workers have different operating proficiency and work ability, so different workers have different processing times for the same process, different workers set different parameters for the machine, and different processing energy consumption; the scheduling goal is to arrange a reasonable processing sequence for each process and select a reasonable combination of processing machines and workers, objective function and energy consumption indicators while satisfying constraints and considering interference events in the workshop;

[0015] The premise assumptions for establishing the flexible job shop scheduling problem include:

[0016] Assume that the start time of workpiece 1 is after time 0;

[0017] Assumption 2: The workpiece needs to be processed strictly according to the process path;

[0018] Assume that the three machines can only process one workpiece at a time and there can be no conflict;

[0019] Assumption 4: Different artifacts are independent of each other;

[0020] The multi-objective flexible job shop scheduling problem under dual resource constraints adds worker constraints on the basis of the flexible job shop scheduling model. The decision variables of the dual resource scheduling model are:

[0021]

[0022]

[0023]

[0024] The mathematical model of the multi-objective flexible job shop scheduling problem under dual resource constraints is as follows:

[0025] s jh +x irjh ×p irjh ≤c jh (1)

[0026] Among them, s jh is the processing start time of process Ojh, x irjh is the decision variable, p ijh It is process O jh Processing time on machine i, c j is the processing completion time of the jth workpiece;

[0027] c jh ≤s j(h+1) (2)

[0028] Among them, c j is the processing completion time of the jth workpiece, s jh It is process O jh The processing start time;

[0029] s jh +p ijh ≤s kl +L(1-y ijhkl ) (3)

[0030] Among them, s jh is the processing start time of process Ojh, p ijh It is process O jh The processing time on machine i, L is an infinite positive number, y ijhkl is the decision variable;

[0031] c jh ≤sj(h+1) +L(1-y iklj(h+1) ) (4)

[0032] Among them, c j is the processing completion time of the jth workpiece, s jh It is process O jh The processing start time, L is an infinite positive number;

[0033] s jh +p irjh ≤s kl +L(1-z rijhkl ) (5)

[0034] Among them, p ijh It is process O jh Processing time on machine i, s jh It is process O jh The processing start time, L is an infinite positive number;

[0035] c jh ≤s j(h+1) +L(1-z riklj(h+1) ) (6)

[0036] Among them, c j is the processing completion time of the jth workpiece, s jh It is process O jh The processing start time, L is an infinite positive number;

[0037]

[0038]

[0039] C j ≤C max (9)

[0040]

[0041] In formula 1-10, i∈{1,2,3,...,m}, j∈{1,2,3,...,n}, h∈{1,2,3,...,h k}, k∈{1,2,3,...,n}, l∈{1,2,3,...,h k}; Formula 1 indicates that each workpiece must be processed according to a fixed process route and cannot be changed; Formula 2 indicates that the completion time of any process cannot exceed the start time; Formula 3 indicates that the same machine can only process one process at the same time; Formula 4 indicates that Formula 5 and Formula 6 indicate that a worker can only operate one machine at the same time; the workpiece cannot be disassembled, Formula 7 indicates that a process can only be processed on one machine at the same time; Formula 8 indicates that a process can only be processed by one worker operating one machine at the same time; Formula 9 indicates that the completion time of the workpiece is less than or equal to the maximum completion time;

[0042] TE=TWE+TPE (11)

[0043] Among them, TE is the total energy consumption of the workshop, TWE is the no-load energy consumption, and TPE is the load energy consumption;

[0044]

[0045]

[0046] Formula 10 indicates that the total energy consumption of the workshop TE is the sum of the no-load energy consumption TWE and the load energy consumption TPE; Formula 12 indicates that worker r operates machine i in process O jh The processing energy consumption; Formula 13 represents the no-load energy consumption of workshop scheduling.

[0047] Specifically, the improved hybrid particle swarm algorithm includes the following processes:

[0048] Initialize the number of iterations Iter, the population size pop, and the value of the inertia constant ω parameter, and use the mixed initialization method to generate the initial population;

[0049] Calculate the fitness value of each particle and sort them, and initialize the global worst solution P w , the global optimal solution P g and the local optimal solution P i ;

[0050] Update particles according to the particle update formula;

[0051] Update P w , P g and P i And select the three particles with the smallest maximum completion time;

[0052] Perform variable neighborhood search operation on the three particles selected in the previous step;

[0053] Update P w , P g and P i ;

[0054] Determine whether the end condition is met. If so, output the optimal individual and the maximum completion cycle. Otherwise, update the particles according to the particle update formula.

[0055] Furthermore, the hybrid particle swarm algorithm is improved by using a collaborative optimization strategy to obtain an improved hybrid particle swarm algorithm, including:

[0056] Input order information, encode processes, machines and workers by particles, and use plug-in decoding to obtain feasible scheduling solutions;

[0057] According to the particle encoding information, the population is initialized using a hybrid initialization method;

[0058] Introduce Jaya algorithm update ideas and crossover and mutation operations to update the scheduling scheme;

[0059] Expand the search space of solutions based on the neighborhood structure and update the scheduling plan;

[0060] Determine whether the iteration conditions are met. If the maximum number of iterations is not reached, update the scheduling plan. If the maximum number of iterations is reached, output the optimal scheduling plan and multiple optimization objective values.

[0061] Furthermore, the encoding and decoding scheme includes:

[0062] Coding refers to the expression form of individuals in a population, and decoding is the process of converting each solution into a corresponding scheduling plan. In the decoding process, process constraints need to be considered. In actual manufacturing, the processing sequence of all processes of the workpiece will be designed in the process design stage, and once fixed, it is rarely changed. Therefore, only when the previous process is completed can the next process be started. At the same time, the constraints of the machine must also be considered. Resources are unique, and a machine can only process one process at a time. The encoding vector, MS machine selection part, is arranged according to the order of the workpiece and process, which is O 11 →O 12 →O 21 →O 22 →O 31 →O 32 →O 33 ; Different numbers in the OS process sorting sequence represent different workpieces, and the number of times the same number appears represents the sequence number of the process.

[0063] Specifically, the hybrid initialization method is used to initialize the population, which means:

[0064] The OS process sequence is generated by random selection. The machine selection part uses a hybrid initialization method that combines global selection, local selection and random selection to generate the initial population. The ratios of the three initialization methods are 0.6, 0.3, and 0.1. The random selection initialization method is used to generate the MS part to expand the search space. The process is as follows: traverse all processes, determine the set of machines to be selected, and randomly select a machine from it and record it to generate the initial machine selection sequence.

[0065] Specifically, the method of expanding the search space of the solution according to the neighborhood structure and updating the scheduling scheme includes:

[0066] Initialize the population and define multiple neighborhood structures, N k ,k=1,...,k max , generate the initial solution S;

[0067] Let k = 1 and repeat the neighborhood search process until k = k max ;

[0068] Explore the neighborhood of solution S to obtain S'; compare solutions S' and S. If S' is better, S = S', k = 1; otherwise k = k + 1.

[0069] Specifically, the step 4 comprises the following steps:

[0070] Step 41, determining whether a machine failure occurs in the workshop, if no machine failure occurs, executing the next step, if a machine failure occurs, determining the faulty machine information;

[0071] Step 42, determining the processing parameters of the new workpiece according to the information of the new workpiece to be processed;

[0072] The process of step 41 includes:

[0073] Step 411, randomly generate a faulty machine i;

[0074] Step 412: The time interval between two consecutive failures of machine i and the repair time MFT are obtained by equations 14 and 15: i and MRT i ;

[0075] MFT i =exp rnd(MTBF) (14)

[0076] MRT i =exprnd(MTTF) (15)

[0077] Step 413: Failure time BT of machine i i , Repair time RT i and available time MT iObtained by equations 16, 17 and 18 respectively;

[0078] BT i =MT i +MFT i (16)

[0079] MT i =BT i +MRT i (17)

[0080] RT i =MRT i (18)

[0081] The process of step 42 includes:

[0082] Step 421, randomly generate the workpiece arrival time;

[0083] Step 422: randomly select a workpiece and add it to the set of workpieces to be processed.

[0084] A multi-objective flexible workshop dynamic scheduling device under dual resource constraints of the present invention comprises:

[0085] The model building module is used to determine the dual resource constraints of machines and workers in the production workshop and the production assumptions; determine the multi-objective optimization indicators, including the maximum completion time and workshop energy consumption, and obtain the objective function; analyze the multi-objective dynamic scheduling characteristics of the dual-resource flexible workshop, and establish a multi-objective static scheduling model for the flexible workshop under dual resource constraints;

[0086] The execution module is used to obtain the processing information of the orders to be processed and the workers' scheduling information from the workshop, and generate a pre-scheduling plan based on the improved hybrid particle swarm algorithm for the workshop production execution;

[0087] The trigger module is used to monitor the real-time production status of the workshop using a dynamic event trigger mechanism to determine whether a disturbance is triggered;

[0088] A parameter module is used to determine dynamic event parameters for two dynamic events: machine failure and new workpiece arrival;

[0089] The decision-making module is used to reschedule the processing tasks using an improved hybrid particle swarm algorithm to obtain the optimal scheduling decision plan, completion cycle and workshop energy consumption under dual resource constraints.

[0090] An electronic device of the present invention comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the multi-objective flexible workshop dynamic scheduling method under dual resource constraints as described above is implemented.

[0091] A non-transitory computer-readable storage medium of the present invention stores a computer program thereon, characterized in that when the computer program is executed by a processor, the multi-objective flexible workshop dynamic scheduling method under dual resource constraints as described above is implemented.

[0092] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0093] 1. The present invention designs an improved hybrid particle swarm algorithm that combines a particle swarm algorithm with a variable neighborhood search algorithm to solve the flexible workshop scheduling problem under dual resource constraints. The hybrid initialization method, hybrid update method and variable neighborhood search operation can effectively shorten the solution time of the scheduling scheme, reduce workshop production costs and optimize resource utilization;

[0094] 2. The present invention studies the dynamic scheduling method of flexible workshops under the dual resource constraints of machines and workers, obtains the multi-objective optimization Pareto scheduling decision solution set of flexible workshops, shortens the production workshop construction period by more than 10%, reduces workshop production energy consumption and improves production efficiency;

[0095] 3. The present invention comprehensively considers two kinds of dynamic interference, namely machine failure and random arrival of workpieces, and adopts an improved hybrid particle swarm algorithm to perform rescheduling on the pre-scheduling plan, thereby reducing the adverse effects of interference events on workshop production, improving the robustness of the scheduling plan, and promoting the transformation and upgrading of industrial production lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Figure 1 The present invention is a flow chart of a flexible workshop scheduling method under dual resource constraints.

[0097] Figure 2 The figure is a schematic diagram of the process of an improved hybrid particle swarm algorithm according to an embodiment of the present invention.

[0098] Figure 3 A schematic diagram of an integer encoding method according to an embodiment of the present invention.

[0099] Figure 4 A decoding Gantt chart according to an embodiment of the present invention.

[0100] Figure 5 This is a schematic diagram of initializing a global selection initialization method according to an embodiment of the present invention.

[0101] Figure 6 This is a schematic diagram of initialization of a local selection initialization method according to an embodiment of the present invention.

[0102] Figure 7a , Figure 7b This is a Gantt chart of a scheduling scheme for a scheduling example of the present invention. Figure 7a A Gantt chart for pre-scheduling workers for a scheduling example of the present invention, Figure 7b A Gantt chart is provided for pre-scheduling machine allocation for a scheduling example of the present invention.

[0103] Figure 8a-8c This is a schematic diagram of the results of a dynamic scheduling solution for scenario 1 of the present invention. Figure 8a A Gantt chart for pre-scheduling workers for a scheduling example of the present invention, Figure 8b A Gantt chart is provided for pre-scheduling machine allocation for a scheduling example of the present invention. Figure 8c It is a schematic diagram of the Pareto solution collection and distribution points of a multi-objective optimization of a dynamic scenario 1 of the present invention.

[0104] Figure 9a-9c This is a schematic diagram of the results of a dynamic scheduling solution for scenario 2 of the present invention. Figure 9a A Gantt chart for pre-scheduling workers for a scheduling example of the present invention, Figure 9b A Gantt chart is provided for pre-scheduling machine allocation for a scheduling example of the present invention. Fig.9c It is a schematic diagram of the Pareto solution collection and distribution points of a multi-objective optimization of a dynamic scenario 2 of the present invention.

[0105] Figure 10a-10c This is a schematic diagram of the results of a dynamic scheduling solution for scenario 3 of the present invention. Fig.10a A Gantt chart for pre-scheduling workers for a scheduling example of the present invention, Fig.10b A Gantt chart is provided for pre-scheduling machine allocation for a scheduling example of the present invention. Fig.10c It is a schematic diagram of the Pareto solution collection and distribution points of a multi-objective optimization of a dynamic scenario 3 of the present invention.

[0106] Figures 11a-11c This is a schematic diagram of the results of a dynamic scheduling solution for scenario 4 of the present invention. Fig.11a A Gantt chart for pre-scheduling workers for a scheduling example of the present invention, Fig.11b A Gantt chart is provided for pre-scheduling machine allocation for a scheduling example of the present invention. Fig.11c It is a schematic diagram of the Pareto solution collection and distribution points of a multi-objective optimization of a dynamic scene 4 of the present invention. DETAILED DESCRIPTION

[0107] The present invention discloses a multi-objective flexible workshop dynamic scheduling method, device and electronic equipment under dual resource constraints, including a mathematical model process for dual resource flexible workshop scheduling, a specific solution process for workshop scheduling scheme using an improved hybrid particle swarm algorithm, and a rescheduling scheme for solving dynamic workshop scheduling problems under complex interference.

[0108] The present invention is further described in detail below with reference to the accompanying drawings.

[0109] like Figure 1As shown, the present invention is a flexible workshop dynamic scheduling method considering the dual resource constraints of machines and workers. First, a static production scheduling model under dual resource constraints is constructed, and then the information of pending orders and worker shifts is obtained from the production workshop. A pre-scheduling scheme is generated based on an improved hybrid particle swarm algorithm. The real-time production status is judged through a dynamic event trigger mechanism. The two types of interference, machine failure and the arrival of new workpieces, are mainly considered, and the machine failure parameters and new workpiece information are determined. The processes and machines in the workshop are classified. Finally, the improved hybrid particle swarm algorithm is used to reschedule the workshop processing tasks to solve the optimal scheduling scheme under the workshop scheduling model under dual resource constraints. Specifically, the following steps are included:

[0110] Step 1: Determine the dual resource constraints of machines and workers in the production workshop and the production assumptions, and establish a multi-objective static scheduling model for the flexible workshop under dual resource constraints;

[0111] Step 2: Obtain the processing information of the orders to be processed and the workers' schedule information from the workshop, and generate a pre-scheduling plan based on the improved hybrid particle swarm algorithm for the workshop production execution;

[0112] Step 3: Use a dynamic event trigger mechanism to monitor the real-time production status of the workshop and determine whether a disturbance is triggered;

[0113] Step 4: Determine dynamic event parameters for two dynamic events: machine failure and new workpiece arrival;

[0114] Step 5: Use the improved hybrid particle swarm algorithm to reschedule the processing tasks and obtain the optimal scheduling decision plan, completion cycle and workshop energy consumption under dual resource constraints.

[0115] The step 1 specifically comprises the following steps:

[0116] Step 11: Determine the premise and assumptions under the dual resource constraints and define the variables based on the description of the shop scheduling problem;

[0117] Step 12: Determine the multi-objective optimization indicators, including the maximum completion time and workshop energy consumption, and obtain the objective function;

[0118] Step 13: Analyze the multi-objective dynamic scheduling characteristics of the dual-resource flexible workshop and establish a workshop scheduling model under multiple constraints.

[0119] The assumptions of the present invention are as follows:

[0120] The description of the multi-objective dual-resource flexible workshop scheduling problem is as follows: w workers operate m machines to complete the processing task of n workpieces. Each workpiece consists of at least one process. Any process can be processed on one or more machines, and each worker can operate one or more machines. Different workers have different operating proficiency and work ability. Therefore, different workers have different processing times for the same process, different workers set different parameters for the machine, and different processing energy consumption. The scheduling goal is to arrange a reasonable processing sequence for each process and select a reasonable combination of processing machines and workers, objective functions and energy consumption indicators while meeting the constraints and considering the interference events in the workshop.

[0121] The premise assumptions for establishing the flexible job shop scheduling problem include:

[0122] Assume that the start time of workpiece 1 is after time 0;

[0123] Assumption 2: The workpiece needs to be processed strictly according to the process path;

[0124] Assume that the three machines can only process one workpiece at a time and there can be no conflict;

[0125] Assumption 4: Different artifacts are independent of each other.

[0126] The multi-objective flexible job shop scheduling problem under dual resource constraints adds worker constraints on the basis of the flexible job shop scheduling model. The decision variables of the dual resource scheduling model are:

[0127]

[0128]

[0129]

[0130] The mathematical model of the multi-objective flexible job shop scheduling problem under dual resource constraints is as follows:

[0131] s jh +x irjh ×p irjh ≤c jh (1)

[0132] Among them, s jh is the processing start time of process Ojh, x irjh is the decision variable, p ijh It is process O jh Processing time on machine i, c j is the processing completion time of the jth workpiece;

[0133] c jh ≤s j(h+1) (2)

[0134] Among them, c j is the processing completion time of the jth workpiece, s jh It is process O jh The processing start time;

[0135] s jh +p ijh ≤s kl +L(1-y ijhkl ) (3)

[0136] Among them, s jh is the processing start time of process Ojh, p ijh It is process O jh The processing time on machine i, L is an infinite positive number, y ijhkl is the decision variable;

[0137] c jh ≤s j(h+1) +L(1-y iklj(h+1) ) (4)

[0138] Among them, c j is the processing completion time of the jth workpiece, s jh It is process O jh The processing start time, L is an infinite positive number.

[0139] s jh +p irjh ≤s kl +L(1-z rijhkl ) (5)

[0140] Among them, p ijh It is process O jh Processing time on machine i, s jh It is process O jh The processing start time, L is an infinite positive number.

[0141] c jh ≤s j(h+1) +L(1-z riklj(h+1) ) (6)

[0142] Among them, c j is the processing completion time of the jth workpiece, s jh It is process O jh The processing start time, L is an infinite positive number.

[0143]

[0144]

[0145] C j ≤C max (9)

[0146]

[0147] In formula 1-10, i∈{1,2,3,...,m}, j∈{1,2,3,...,n}, h∈{1,2,3,...,h k}, k∈{1,2,3,...,n}, l∈{1,2,3,...,h k Formula 1 indicates that each workpiece must be processed according to a fixed process route and cannot be changed; Formula 2 indicates that the completion time of any process cannot exceed the start time; Formula 3 indicates that the same machine can only process one process at the same time; Formula 4 indicates that Formula 5 and Formula 6 indicate that a worker can only operate one machine at the same time; the workpiece cannot be disassembled, Formula 7 indicates that a process can only be processed on one machine at the same time; Formula 8 indicates that a process can only be processed by one worker operating one machine at the same time; Formula 9 indicates that the completion time of the workpiece is less than or equal to the maximum completion time;

[0148] TE=TWE+TPE (11)

[0149] Among them, TE is the total energy consumption of the workshop, TWE is the no-load energy consumption, and TPE is the load energy consumption;

[0150]

[0151]

[0152] Formula 10 indicates that the total energy consumption of the workshop TE is the sum of the no-load energy consumption TWE and the load energy consumption TPE; Formula 12 indicates that worker r operates machine i in process O jh The processing energy consumption; Formula 13 represents the no-load energy consumption of workshop scheduling.

[0153] The step 2 specifically includes the following steps:

[0154] Step 21: Obtain model input information from the workshop, including the processing machines, processing time, processing power and worker information that can operate the machine tools for different workpieces;

[0155] Step 22: Determine the multi-objective optimization indicators, including the maximum completion time and the energy consumption of the workshop, and obtain the objective function;

[0156] Step 23: Use collaborative optimization strategy to improve the hybrid particle swarm algorithm to obtain an improved hybrid particle swarm algorithm.

[0157] like Figure 2 The improved hybrid particle swarm algorithm shown includes the following processes:

[0158] Initialize the values ​​of parameters such as the number of iterations Iter, the population size pop, the inertia constant ω, etc., and use the mixed initialization method to generate the initial population;

[0159] Calculate the fitness value of each particle and sort them, and initialize the global worst solution P w 、Global optimal solution P g and the local optimal solution P i ;

[0160] Update particles according to the particle update formula;

[0161] Update P w , P g and P i And select the three particles with the smallest maximum completion time;

[0162] Perform variable neighborhood search operation on the three particles selected in the previous step;

[0163] Update P w , P g and P i ;

[0164] Determine whether the end condition is met. If it is met, output the optimal individual and the maximum completion cycle. Otherwise, update the particles according to the particle update formula and execute the loop operation.

[0165] The step 23 specifically includes the following steps:

[0166] Step 231: input order information, perform particle encoding on processes, machines and workers, and use plug-in decoding to obtain a feasible scheduling solution;

[0167] Step 232: Initialize the population using a hybrid initialization method according to the particle encoding information;

[0168] Step 233: Introduce the Jaya algorithm update idea and crossover and mutation operations to update the scheduling plan;

[0169] Step 234: Neighborhood search is a method for finding a better solution by adjusting the local structure of the solution, expanding the search space of the solution according to the neighborhood structure, and updating the scheduling plan;

[0170] Step 235: Determine whether the iteration condition is met. If the maximum number of iterations is not reached, return to step 233. If the maximum number of iterations is reached, output the optimal scheduling plan and multiple optimization target values.

[0171] The encoding and decoding scheme described in step 231 is as follows:

[0172] Encoding refers to the expression form of individuals in a population, and decoding is the process of converting each solution into a corresponding scheduling plan. The decoding method has a crucial impact on the quality of the scheduling plan. In the decoding process, process constraints need to be considered. In actual manufacturing, the processing sequence of all processes of the workpiece will be designed in the process design stage. Once fixed, it is rarely changed. Therefore, the next process can only be started after the previous process is completed. At the same time, the constraints of the machine must also be considered. Resources are unique, and a machine can only process one process at a time.

[0173] Table 1 FJSP examples for three machines and three workpieces

[0174]

[0175] like Figure 3 The code vector corresponding to the FJSP example in Table 1 is shown in Table 1. The MS machine selection part is arranged according to the order of the workpiece and the process. 11 →O 12 →O 21 →O 22 →O 31 →O 32 →O 33 For example, the first gene "3" of MS means that the machine with serial number 3 in the available machine set is used to complete the processing task of process O11. Different numbers in the process ranking part of OS represent different workpieces, and the number of times the same number appears represents the sequence number of the process.

[0176] by Figure 4 Take the particle in as an example and use the common decoding method to decode. First, read the OS process sorting sequence of the particle: 21 →O 31 →O 11 →O 22 →O 32 →O 23 →O 12 ; Then, read the machine selection status in the MS machine selection sequence. The machines corresponding to the above processes are: machine 2, machine 2, machine 3, machine 2, machine 3, machine 1, machine 3; finally, read the processing time corresponding to the machine, and then process these processes in sequence. Figure 5 This is the Gantt chart decoded by the common decoding method. The scheduling plan contains 3 workpieces and 7 processes, and the corresponding maximum completion time is 20. When decoding, process O 21 According to the order in OS, it is placed in process O 32 For the subsequent processing, machine 3 is idle in the time period [4,10] and is not fully utilized in this scheduling scheme.

[0177] The hybrid initialization method described in step 232 is as follows:

[0178] The initial population is the beginning of the search. The quality of the initial population has a huge impact on the subsequent iterative process. A high-quality initial population can improve the convergence speed of the solution method. The present invention designs a hybrid initialization method to generate a high-quality initial population. In order to expand the search space, the OS process sorting sequence is generated by a random selection method. The MS machine selection part uses a hybrid initialization method that combines global selection, local selection and random selection to generate the initial population. The ratios of the three initialization methods are: 0.6, 0.3, and 0.1. The random selection initialization method is used to generate the MS part to expand the search space. The specific process of this initialization method is: traverse all processes, determine the set of machines to be selected, randomly select a machine from it and record it, and generate an initial machine selection sequence.

[0179] Table 2 FJSP examples for three machines and two workpieces

[0180]

[0181] In order to ensure the load balance of the machine and improve the utilization of the machine, a hybrid initialization method is introduced. Figure 5 As shown, the specific process of the mixed initialization method is as follows:

[0182] Step 2321: Set an array with a length equal to the number of machines to record the cumulative load of the machines;

[0183] Step 2322: First, randomly select a workpiece from the pool of workpieces to be processed, then traverse each process in order and determine the set of candidate machines for each process, add the processing time of the machines in the set of candidate machines and the cumulative processing time of the machines, select the machine with the shortest time after the addition and update the cumulative load of the machine, and if the processing machines for each process of this workpiece have been determined, execute the next step;

[0184] Step 2323: Delete the workpiece selected in the previous step from the list and execute the operation of the previous step again. Stop the operation when the workpiece list is empty. In order to better explain the initialization method, it will be initialized according to the FJSP example in Table 2. The local search initialization method is very similar to the global search initialization method. The difference is that the local search initialization method does not require random selection of workpieces, but only needs to traverse the workpieces in the order of the workpieces in the workpiece set. In addition, when the process of each workpiece is selected in the local search, the accumulated processing time of the machine should be set to zero to ensure that in the processing process of the same workpiece, the machine with the highest processing efficiency is given priority. The specific process is as follows:

[0185] Step 23231: Set an array with a length equal to the number of machines to record the cumulative load of the machines;

[0186] Step 23232: Take out the workpieces in the order in which they are in the workpiece set, traverse each process of this workpiece in order, determine the set of machines to be selected for each process, add the processing time of the optional machines and the cumulative processing time of the machines, select the machine with the shortest added time and update the cumulative processing time of each machine, if the processing machines for each process of this workpiece have been determined, execute the next step;

[0187] Step 23233: Delete the workpiece selected in the previous step from the list, reset the array recording the accumulated load of the machine to zero and execute the operation of the previous step again. If the workpiece list is empty, stop the operation. Figure 6 The specific process of the local search initialization method.

[0188] The variable neighborhood search operation in step 234 is as follows:

[0189] The improved hybrid particle swarm algorithm expands the search space of the solution by designing a variety of effective neighborhood structures to expand the search range and enhance the local search capability of the scheduling method. The critical path refers to the path with the greatest influence in the scheduling problem. Only by changing the adjacent paths in the key blocks of the critical path can the maximum completion cycle be effectively shortened. The specific process is as follows:

[0190] Step 2341: Initialize the population and define multiple neighborhood structures, N k ,k=1,...,k max , generate the initial solution S;

[0191] Step 2342: Set k = 1 and repeat the neighborhood search process until k = k max ;

[0192] Step 2343: Explore the neighborhood of solution S to obtain S'; compare solutions S' and S, if S' is better, S = S', k = 1; otherwise k = k + 1.

[0193] For the FJSP problem, the design of the neighborhood structure is critical and has the following properties:

[0194] Property 1: Exchanges between non-critical processes cannot shorten the maximum completion time.

[0195] Property 2: The exchange of adjacent processes in the critical block of the critical path will not produce an infeasible solution.

[0196] Property 3: Exchanging adjacent processes that are not the beginning or end of a critical block cannot shorten the maximum completion time.

[0197] In step 3, it is determined whether dynamic interference occurs in the workshop:

[0198] Obtain the workshop pre-scheduling plan to execute the production plan, and continuously monitor the workshop production status, including alarm information, quality inspection, etc. When a machine failure occurs and a new workpiece arrives, transmit the interference information status.

[0199] In step 4, dynamic event parameters are determined:

[0200] The present invention performs rescheduling for the two most common interference events in the workshop: machine failure and new workpiece arrival. When interference occurs, the relevant parameters of the interference event need to be determined. First, a faulty machine i is randomly generated; the time interval and repair time MFT between two adjacent faults of machine i are obtained by equations 14 and 15. i and MRT i ; Failure time BT of machine i i , Repair time RT i and available time MT i They are obtained by equations 16, 17 and 18 respectively. When a new workpiece arrives, it is necessary to determine the arrival time of the randomly generated workpiece and add the randomly selected workpiece to the set of workpieces to be processed.

[0201] The step 4 specifically comprises the following steps:

[0202] Step 41: Determine whether a machine failure occurs in the workshop. If no machine failure occurs, execute step 42. If a machine failure occurs, determine the faulty machine information.

[0203] Step 42: Determine the processing parameters of the new workpiece according to the information of the new workpiece to be processed.

[0204] The step 41 specifically includes the following steps:

[0205] Step 411: Randomly generate a faulty machine i;

[0206] Step 412: The time interval between two consecutive failures of machine i and the repair time MFT are obtained by equations 14 and 15: i and MRT i ;

[0207] MFT i =exprnd(MTBF) (14)

[0208] MRT i =exprnd(MTTF) (15)

[0209] Step 413: Failure time BT of machine i i , Repair time RT i and available time MT i They are obtained by equations 16, 17 and 18 respectively.

[0210] BT i=MT i +MFT i (16)

[0211] MT i =BT i +MRT i (17)

[0212] RT i =MRT i (18)

[0213] The step 42 specifically includes the following steps:

[0214] Step 421: randomly generate workpiece arrival time;

[0215] Step 422: Randomly select a workpiece and add it to the set of workpieces to be processed.

[0216] As described above, the present invention first converts the scheduling decision problem that considers the dual resource constraints of machines and workers into a multi-objective optimization mathematical model problem of workshop scheduling, and then takes the maximum completion time of the order and the energy consumption of the workshop as the optimization target, and considers the constraints such as machine, process priority constraints, and worker resources at the same time, and constructs a flexible workshop multi-objective optimization scheduling model under dual resource constraints; finally, an improved hybrid particle swarm algorithm is used for solving. In the solving process, an improved hybrid particle swarm algorithm combining a particle swarm algorithm and a variable neighborhood search algorithm is designed, and a hybrid initialization method based on combined initialization, a hybrid update method, and a variable neighborhood search operation are used to shorten the solution time of the scheduling scheme. The scheduling scheme obtained by the embodiment of the present invention can effectively solve the flexible workshop scheduling problem with dual resource constraints.

[0217] Step 5: Use the improved hybrid particle swarm algorithm to reschedule the scheduling plan based on the pre-scheduling.

[0218] Partial rescheduling and full rescheduling are two methods that are usually used for rescheduling. Partial rescheduling only reschedules the processes that are directly or indirectly affected by the interference event, without changing the processing plan of other processes. Full rescheduling means that all processes to be processed are rescheduled when a dynamic event occurs. The partial rescheduling method makes little change to the pre-scheduling plan, but the processing cycle is too long, while the full rescheduling method is the opposite. In order to ensure that the processing task can be delivered on time, the present invention uses the full rescheduling method to generate a rescheduling plan.

[0219] The step 5 specifically comprises the following steps:

[0220] Step 51: For two disturbance events, namely, random failure and random arrival of workpieces, an event-driven rescheduling strategy is adopted;

[0221] Step 52: For the two disturbance events of random failure and random arrival of workpieces, a complete rescheduling method is used to generate a rescheduling plan.

[0222] This embodiment is based on the above-mentioned implementation case 1. Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 As shown, a specific experimental verification is used to explain the specific operations of steps 1 to 4 in detail.

[0223] Experimental verification:

[0224] Take the impeller production line of a certain research institute of the aviation industry as an example. This production line mainly produces impellers, mainly including four types of impellers, which can be processed on 6 different CNC machine tools. The processing procedures include rough milling, fine milling, side edge milling, edge milling, fillet milling, point milling and other operations. The correspondence between workers and machines is shown in Table 3, where each machine can be operated by two workers and each worker can operate three machines. Table 4 gives the processing time of the process, and the time required for the most efficient combination of workers and machines to process this process.

[0225] Table 3 Correspondence between operators and machines

[0226]

[0227] Table 4 Standard processing schedule for each process

[0228]

[0229]

[0230] 1. Production task analysis

[0231] The impeller production line involves multiple processes, and there are multiple processable machines in the workshop. The processing time of different processes on the processable machine set is shown in Table 4. Analysis shows that the impeller production line is highly flexible and meets the characteristics of a flexible operation workshop. According to the survey, the production scheduling process of the production workshop mostly relies on manual experience to determine the production sequence. The delivery date of the order and the overall utilization of resources are not standardized and irregular, and order delays and resource waste often occur, and even lead to a decrease in the processing accuracy of parts. At the same time, the equipment in the production line needs to be maintained regularly, and the pre-scheduling plan needs to be updated when a machine failure occurs. The manual scheduling experience exposes the shortcomings of inefficiency and labor, which seriously affects the orderly production in the workshop, causing the order to be delayed and the production efficiency of the enterprise to be reduced. In addition, new orders frequently arrive on the production line. The original production scheduling plan of the workshop is determined by the workers to insert the batch of orders. The insertion time is too late, resulting in the inability to process emergency orders in time. The insertion time is too early, which seriously affects the processing plan of the original order and causes parts rework.

[0232] As shown in Table 3, the workshop is equipped with 4 operators, who can operate different CNC machine tools respectively, so the worker resources are also flexible. The worker resources in the workshop are limited, and the types of work required to operate different processes in different time periods are also different. For example, the type of work required when the impeller is rough milled on machine 1 is W1, and the type of work required when the impeller is fine milled on machine 2 is W2. Conventional scheduling strategies are insufficient for overall control of the workshop, often resulting in staff absence and overtime, which seriously affects production progress.

[0233] 2. Model solution and analysis

[0234] For the above impeller production line processing data, the dual resource constraints of machines and workers need to be considered in the actual scheduling process, that is, the dual resource flexible job shop scheduling model proposed by the present invention. The method proposed by the present invention can provide a pre-scheduling decision plan set and a re-scheduling decision plan set under interference for the production workshop. The present invention uses a software tool based on python 3.8 for simulation optimization and solution, and the parameters of the improved hybrid particle swarm algorithm are set as follows: the maximum number of iterations is set to 100, the population size is 100, the mutation probability is set to 0.15, the acceleration factor and the update probability are 0.5 and 0.7 respectively, and each test problem is run independently 20 times.

[0235] 3. Scheduling example simulation and analysis

[0236] Firstly, the improved hybrid particle swarm algorithm proposed in the present invention is used to solve the scheduling problem model. Under the constraints of machines, workers and complex processes, 20 consecutive calculations are performed. To verify the effectiveness of the improved hybrid particle swarm algorithm proposed in the present invention, the method proposed in the present invention is compared with the non-dominated sorting genetic algorithm and the Djaya algorithm, and the number of non-dominated solutions and the multi-objective optimization indicators in the scheduling solution set are compared.

[0237] Table 5 gives the pre-scheduling results of the three methods, mainly including the number of non-dominated solutions and the Pareto solution with the minimum maximum completion. HPSO obtained 12 non-dominated solutions for the scheduling instance, NSGAII obtained 8 non-dominated solutions, and DJaya obtained 4 non-dominated solutions. HPSO obtained the largest number of non-dominated solutions. [61.0,2252.8] can dominate the Pareto solution with the minimum maximum completion of the other two solutions. Figure 7 is the Gantt chart of the scheduling scheme of the scheduling instance. The scheduling scheme corresponding to the Pareto solution with the minimum maximum completion of HPSO [61.0,2252.8] is the pre-scheduling scheme.

[0238] Table 5 Pre-scheduling results of scheduling examples using three methods

[0239]

[0240] The scheduling scheme in Figure 7 is used as the pre-scheduling scheme for dynamic scheduling. Assume that the machine failure parameters are: in scenario 1, the faulty machine is machine 4, MFT4 and MRT4 are 10 and 20 respectively; in scenario 2, the faulty machine is machine 6, MFT6 and MRT6 are 30 and 20 respectively. The scheduling results of the three methods in the two scenarios are shown in Table 6. Figure 8 is a schematic diagram of the dynamic scheduling scheme results for scenario 1.

[0241] Table 6 The optimal scheduling solutions and the number of non-dominated solutions of the three methods under machine failure

[0242]

[0243] Assume that the arrival parameters of the new workpiece are: workpiece 2 arrives at the 10th time in dynamic scenario 3; workpiece 2 arrives at the 30th time in scenario 4. The scheduling results of the three methods in the two scenarios are shown in Table 7. Figures 8 and 9 are schematic diagrams of the dynamic scheduling scheme results for scenario 1 and scenario 2, respectively.

[0244] Table 7 The optimal scheduling solutions and the number of non-dominated solutions of the three methods under random arrival of workpieces

[0245]

[0246] Figures 10 and 11 are schematic diagrams of the dynamic scheduling results of scenario 3 and scenario 4, respectively. In scenario 3, HPSO obtained 21 non-dominated solutions, NSGAII obtained 2 non-dominated solutions, and DJaya had no non-dominated solutions when solving the scheduling instance. In scenario 4, HPSO obtained 21 non-dominated solutions, NSGAII obtained 2, and DJaya obtained 0. HPSO obtained the largest number of non-dominated solutions and had a higher search accuracy, which proves the effectiveness of the algorithm of the present invention in solving the dynamic scheduling problem.

Claims

1. A multi-objective flexible workshop dynamic scheduling method under dual resource constraints, characterized in that: The following steps are involved: Step 1: Determine the dual resource constraints of machines and workers in the production workshop and the production assumptions; determine the multi-objective optimization indicators, including the maximum completion time and workshop energy consumption, and obtain the objective function; analyze the multi-objective dynamic scheduling characteristics of the dual-resource flexible workshop, and establish a multi-objective static scheduling model for the flexible workshop under dual resource constraints; Step 2: Obtain the processing information of the orders to be processed and the worker scheduling information from the workshop, and generate a pre-scheduling plan based on the improved hybrid particle swarm algorithm for the workshop production execution. The specific process includes: obtaining the input information of the flexible workshop multi-objective static scheduling model from the workshop, including the processing machines, processing time, processing power and worker information of different workpieces that can operate the processing machines; determining the multi-objective optimization indicators, including the maximum completion time and workshop energy consumption, to obtain the objective function; using the collaborative optimization strategy to improve the hybrid particle swarm algorithm to obtain the improved hybrid particle swarm algorithm; Step 3: Use a dynamic event trigger mechanism to monitor the real-time production status of the workshop and determine whether a disturbance is triggered; Step 4: Determine dynamic event parameters for two dynamic events: machine failure and new workpiece arrival; Step 5: Use the improved hybrid particle swarm algorithm to reschedule the processing tasks and obtain the optimal scheduling decision plan, completion cycle and workshop energy consumption under dual resource constraints; The dual resource constraints of production workshop machines and workers and production assumptions include: Assumptions: The multi-objective dual-resource flexible workshop scheduling problem is described as follows: w workers operate m machines to complete the processing task of n workpieces, each workpiece consists of at least one process, any process is processed on one or more machines, and each worker operates one or more machines; different workers have different operating proficiency and work ability, so different workers have different processing times for the same process, different workers set different parameters for the machine, and different processing energy consumption; the scheduling goal is to arrange a reasonable processing sequence for each process and select a reasonable combination of processing machines and workers, objective function and energy consumption indicators while satisfying constraints and considering interference events in the workshop; The premise assumptions for establishing the flexible job shop scheduling problem include: Assume that the start time of workpiece 1 is after time 0; Assumption 2: The workpiece needs to be processed strictly according to the process path; Assume that the three machines can only process one workpiece at a time and there can be no conflict; Assumption 4: Different artifacts are independent of each other; The multi-objective flexible job shop scheduling problem under dual resource constraints adds worker constraints on the basis of the flexible job shop scheduling model. The decision variables of the dual resource scheduling model are: The mathematical model of the multi-objective flexible job shop scheduling problem under dual resource constraints is as follows: s jh +x irjh ×p irjh ≤c jh (1) Among them, s jh It is process O jh The processing start time, x irjh is the decision variable, p irjh It is process O jh Processing time on machine i, c jh is the processing completion time of the jth workpiece; c jh ≤s j(h+1) (2) Among them, c jh is the processing completion time of the jth workpiece, s j(h+1) It is process O j(h+1) The processing start time; s jh +p ijh ≤s kl +L(1-y ijhkl ) (3) Among them, s jh It is process O jh The processing start time, p ijh It is process O jh Processing time on machine i, s kl It is process O jh The processing start time, L is an infinite positive number, y ijhkl is the decision variable; c jh ≤s j(h+1) +L(1-y iklj(h+1) ) (4) Among them, c jh is the processing completion time of the jth workpiece, s j(h+1) It is process O j(h+1) The processing start time, L is an infinite positive number, y iklj(h+1) is the decision variable; s jh +p irjh ≤s kl +L(1-z rijhkl ) (5) Among them, s jh It is process O jh The processing start time, p irjh It is process O jh Processing time on machine i, s kl It is process O kl The processing start time, L is an infinite positive number, z rijhkl is the decision variable; c jh ≤s j(h+1) +L(1-z riklj(h+1) ) (6) Among them, c jh is the processing completion time of the jth workpiece, s j(h+1) It is process O j(h+1) The processing start time, L is an infinite positive number, z riklj(h+1) is the decision variable; C j ≤C max (9) In formula 1-10, i∈{1,2,3,...,m}, j∈{1,2,3,...,n}, h∈{1,2,3,...,h k }, k∈{1,2,3,...,n}, l∈{1,2,3,...,h k }; Formula 1 indicates that each workpiece must be processed according to a fixed process route and cannot be changed; Formula 2 indicates that the completion time of any process cannot exceed the start time; Formula 3 indicates that the same machine can only process one process at the same time; Formula 4 indicates that Formula 5 and Formula 6 indicate that a worker can only operate one machine at the same time; the workpiece cannot be disassembled, Formula 7 indicates that a process can only be processed on one machine at the same time; Formula 8 indicates that a process can only be processed by one worker operating one machine at the same time; Formula 9 indicates that the completion time of the workpiece is less than or equal to the maximum completion time; Formula 10 indicates the minimized maximum completion time of the jth workpiece; TE=TWE+TPE (11) Among them, TE is the total energy consumption of the workshop, TWE is the no-load energy consumption, and TPE is the load energy consumption; Formula 11 indicates that the total energy consumption of the workshop TE is the sum of the no-load energy consumption TWE and the load energy consumption TPE; Formula 12 indicates that worker r operates machine i in process O jh The processing energy consumption; Formula 13 represents the no-load energy consumption of workshop scheduling.

2. The multi-objective flexible workshop dynamic scheduling method under dual resource constraints according to claim 1 is characterized in that: The improved hybrid particle swarm algorithm, The process includes: Initialize the number of iterations Iter, the population size pop, and the value of the inertia constant ω parameter, and use the mixed initialization method to generate the initial population; Calculate the fitness value of each particle and sort them, and initialize the global worst solution P w 、Global optimal solution P g and the local optimal solution P i ; Update particles according to the particle update formula; Update P w , P g and P i And select the three particles with the smallest maximum completion time; Perform variable neighborhood search operation on the three particles selected in the previous step; Update P w , P g and P i ; Determine whether the end condition is met. If so, output the optimal individual and the maximum completion cycle. Otherwise, update the particles according to the particle update formula.

3. The multi-objective flexible workshop dynamic scheduling method under dual resource constraints according to claim 1 is characterized in that: The improved hybrid particle swarm algorithm by adopting the collaborative optimization strategy is obtained, and includes: Input order information, encode the process, machine and worker by particles, and use plug-in decoding to obtain the feasible solution for scheduling; According to the particle encoding information, the population is initialized using a hybrid initialization method; Introduce Jaya algorithm update ideas and crossover and mutation operations to update the scheduling scheme; Expand the search space of solutions based on the neighborhood structure and update the scheduling plan; Determine whether the iteration conditions are met. If the maximum number of iterations is not reached, update the scheduling plan. If the maximum number of iterations is reached, output the optimal scheduling plan and multiple optimization objective values.

4. The multi-objective flexible workshop dynamic scheduling method under dual resource constraints according to claim 3 is characterized in that: The encoding and decoding schemes include: Coding refers to the expression form of individuals in the population, and decoding is the process of converting each solution into a corresponding scheduling plan. In the decoding process, process constraints need to be considered. In actual manufacturing, the processing sequence of all processes of the workpiece will be designed in the process design stage, and once fixed, it is rarely changed. Therefore, only when the previous process is completed can the next process be started. At the same time, the constraints of the machine must also be considered. Resources are unique, and a machine can only process one process at a time. The encoding vector, the MS machine selection part, is arranged according to the order of the workpiece and the process, which is O 11 →O 12 →O 21 →O 22 →O 31 →O 32 →O 33 ; Different numbers in the OS process sorting sequence represent different workpieces, and the number of times the same number appears represents the sequence number of the process.

5. The multi-objective flexible workshop dynamic scheduling method under dual resource constraints according to claim 3 is characterized in that: The hybrid initialization method is used to initialize the population, which means: The OS process sequence is generated by random selection. The machine selection part uses a hybrid initialization method that combines global selection, local selection and random selection to generate the initial population. The ratios of the three initialization methods are 0.6, 0.3 and 0.

1. The MS part is generated by random selection initialization method to expand the search space. The process is as follows: traverse all processes, determine the set of machines to be selected and randomly select a machine from them and record it to generate the initial machine selection sequence.

6. The multi-objective flexible workshop dynamic scheduling method under dual resource constraints according to claim 3 is characterized in that: The method of expanding the search space of the solution according to the neighborhood structure and updating the scheduling scheme includes: Initialize the population and define multiple neighborhood structures, N k ,k=1,...,k max , generate the initial solution S; Let k = 1 and repeat the neighborhood search process until k = k max ; Explore the neighborhood of solution S to obtain S'; compare solutions S' and S. If S' is better, S = S', k = 1; otherwise k = k + 1.

7. The multi-objective flexible workshop dynamic scheduling method under dual resource constraints according to claim 1 is characterized in that: The step 4 comprises the following steps: Step 41, determining whether a machine failure occurs in the workshop, if no machine failure occurs, executing the next step, if a machine failure occurs, determining the faulty machine information; Step 42, determining the processing parameters of the new workpiece according to the information of the new workpiece to be processed; The process of step 41 includes: Step 411, randomly generate a faulty machine i; Step 412: The time interval between two consecutive failures of machine i and the repair time MFT are obtained by equations 14 and 15: i and MRT i ; MFT i =exprnd(MTBF) (14) MRT i =exprnd(MTTF) (15) Step 413: Failure time BT of machine i i , Repair time RT i and available time MT i Obtained by equations 16, 17 and 18 respectively; BT i =MT i +MFT i (16) MT i =BT i +MRT i (17) RT i =MRT i (18) The process of step 42 includes: Step 421, randomly generate the workpiece arrival time; Step 422: randomly select a workpiece and add it to the set of workpieces to be processed.

8. A multi-objective flexible workshop dynamic scheduling device under dual resource constraints, characterized in that: include: Model building module, used to determine the dual resource constraints of machines and workers in the production workshop and production assumptions; Determine the multi-objective optimization indicators, including the maximum completion time and workshop energy consumption, and obtain the objective function; analyze the multi-objective dynamic scheduling characteristics of the dual-resource flexible workshop, and establish a multi-objective static scheduling model for the flexible workshop under dual-resource constraints; The execution module is used to obtain the processing information of the orders to be processed and the workers' scheduling information from the workshop, and generate a pre-scheduling plan based on the improved hybrid particle swarm algorithm for the workshop production execution; The trigger module is used to monitor the real-time production status of the workshop using a dynamic event trigger mechanism to determine whether a disturbance is triggered; A parameter module is used to determine dynamic event parameters for two dynamic events: machine failure and new workpiece arrival; The decision-making module is used to reschedule the processing tasks using an improved hybrid particle swarm algorithm to obtain the optimal scheduling decision plan, completion cycle and workshop energy consumption under dual resource constraints; The dual resource constraints of production workshop machines and workers and production assumptions include: Assumptions: The multi-objective dual-resource flexible workshop scheduling problem is described as follows: w workers operate m machines to complete the processing task of n workpieces, each workpiece consists of at least one process, any process is processed on one or more machines, and each worker operates one or more machines; different workers have different operating proficiency and work ability, so different workers have different processing times for the same process, different workers set different parameters for the machine, and different processing energy consumption; the scheduling goal is to arrange a reasonable processing sequence for each process and select a reasonable combination of processing machines and workers, objective function and energy consumption indicators while satisfying constraints and considering interference events in the workshop; The premise assumptions for establishing the flexible job shop scheduling problem include: Assume that the start time of workpiece 1 is after time 0; Assumption 2: The workpiece needs to be processed strictly according to the process path; Assume that the three machines can only process one workpiece at a time and there can be no conflict; Assumption 4: Different artifacts are independent of each other; The multi-objective flexible job shop scheduling problem under dual resource constraints adds worker constraints on the basis of the flexible job shop scheduling model. The decision variables of the dual resource scheduling model are: The mathematical model of the multi-objective flexible job shop scheduling problem under dual resource constraints is as follows: s jh +x irjh ×p irjh ≤c jh (1) Among them, s jh It is process O jh The processing start time, x irjh is the decision variable, p irjh It is process O jh Processing time on machine i, c jh is the processing completion time of the jth workpiece; c jh ≤s j(h+1) (2) Among them, c jh is the processing completion time of the jth workpiece, s j(h+1) It is process O j(h+1) The processing start time; s jh +p ijh ≤s kl +L(1-y ijhkl ) (3) Among them, s jh It is process O jh The processing start time, p ijh It is process O jh Processing time on machine i, s kl It is process O jh The processing start time, L is an infinite positive number, y ijhkl is the decision variable; c jh ≤s j(h+1) +L(1-y iklj(h+1) ) (4) Among them, c jh is the processing completion time of the jth workpiece, s j(h+1) It is process O j(h+1) The processing start time, L is an infinite positive number, y iklj(h+1) is the decision variable; s jh +p irjh ≤s kl +L(1-z rijhkl ) (5) Among them, s jh It is process O jh The processing start time, p irjh It is process O jh Processing time on machine i, s kl It is process O kl The processing start time, L is an infinite positive number, z rijhkl is the decision variable; c jh ≤s j(h+1) +L(1-z riklj(h+1) ) (6) Among them, c jh is the processing completion time of the jth workpiece, s j(h+1) It is process O j(h+1) The processing start time, L is an infinite positive number, z riklj(h+1) is the decision variable; C j ≤C max (9) In formula 1-10, i∈{1,2,3,...,m}, j∈{1,2,3,...,n}, h∈{1,2,3,...,h k }, k∈{1,2,3,...,n}, l∈{1,2,3,...,h k }; Formula 1 indicates that each workpiece must be processed according to a fixed process route and cannot be changed; Formula 2 indicates that the completion time of any process cannot exceed the start time; Formula 3 indicates that the same machine can only process one process at the same time; Formula 4 indicates that Formula 5 and Formula 6 indicate that a worker can only operate one machine at the same time; the workpiece cannot be disassembled, Formula 7 indicates that a process can only be processed on one machine at the same time; Formula 8 indicates that a process can only be processed by one worker operating one machine at the same time; Formula 9 indicates that the completion time of the workpiece is less than or equal to the maximum completion time; Formula 10 indicates the minimized maximum completion time of the jth workpiece; TE=TWE+TPE (11) Among them, TE is the total energy consumption of the workshop, TWE is the no-load energy consumption, and TPE is the load energy consumption; Formula 11 indicates that the total energy consumption of the workshop TE is the sum of the no-load energy consumption TWE and the load energy consumption TPE; Formula 12 indicates that worker r operates machine i in process O jh The processing energy consumption; Formula 13 represents the no-load energy consumption of workshop scheduling.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the multi-objective flexible workshop dynamic scheduling method under dual resource constraints as described in any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the multi-objective flexible workshop dynamic scheduling method under dual resource constraints as claimed in any one of claims 1 to 7 is implemented.

Images