Flexible job shop scheduling optimization method considering machine pre-maintenance and man-hour controllability

Abstract

The invention discloses a flexible job shop scheduling optimization method considering machine pre-maintenance and man-hour controllability. The method comprises the following steps: analyzing the influence of a machine tool pre-maintenance and shutdown / restart strategy on the energy consumption of a flexible job shop, and carrying out modeling; performing optimization solution on the model by adopting an MOPSO and NSGA-II hybrid algorithm, including initializing parameters, generating an initial population by adopting a roulette initialization mode, performing non-dominated sorting and selection operation on the population, updating particle speed and position and the like; validity of the model and the method is verified through examples, wherein configuration information of a workpiece to be machined and a workshop is given, and optimization result analysis is carried out. The method is suitable for flexible job shop scheduling with the problems of machine tool equipment failure, machine idle waiting and the like, the production efficiency can be effectively improved, the energy consumption is reduced, and the stability and the reliability of shop scheduling are improved.

Description

Technical Field

[0001] The invention relates to a flexible job shop scheduling method, in particular to a flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours. Background Art

[0002] As the main body of the equipment manufacturing industry, manufacturing equipment is irreplaceable equipment in all fields of the manufacturing industry. It has many energy sources, complex energy consumption characteristics, and large total energy consumption. It is the main energy consumption in the manufacturing process. In the flexible workshop, the operation of manufacturing equipment including CNC machine tools directly affects the production efficiency and energy consumption level.

[0003] In recent years, many scholars have conducted various studies on the flexible job shop scheduling problem, but most of the existing studies only consider minimizing the maximum completion time as the single-objective FJSP solution, ignoring the energy consumption in the production process. With the increasing severity of environmental problems and the development of green manufacturing, it is becoming increasingly important to consider energy consumption-related goals in production scheduling, but these studies all assume that machines are continuously available.

[0004] In fact, the scheduling model that considers machine availability constraints is more in line with the actual production environment and helps to improve the stability of the production plan. In this context, the integrated optimization research of production scheduling and pre-maintenance planning has gradually attracted attention. In the actual processing process, machine tool equipment failure will lead to processing interruption, increased energy consumption and affected scheduling. Machine idle waiting will also have a negative impact. Therefore, it is very important to consider the pre-maintenance of machine tools and when to turn on and off the machines in production scheduling. However, the current relevant research results are very limited, which makes the workshop scheduling plan unable to be close to the actual production situation, and it is difficult to overcome the negative impact of machine tool failure and machine tool idle waiting, resulting in low processing efficiency and more energy consumption. Summary of the invention

[0005] Purpose of the invention: In view of the above problems, the purpose of the present invention is to provide a flexible job shop scheduling optimization method that takes into account machine preventive maintenance and controllable working hours, which is suitable for flexible job shop scheduling, so that the shop scheduling plan is more in line with reality, improves productivity and reduces energy consumption.

[0006] Technical solution: A flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours includes the following steps:

[0007] S1: Analyze the impact of machine tool pre-maintenance, machine tool shutdown and restart strategies on the energy consumption of the flexible job shop, and model the FJSP considering machine tool pre-maintenance, machine tool shutdown and restart strategies. The model includes objective function and constraint model.

[0008] S2: Use the MOPSO and NSGA-II hybrid algorithm to optimize and solve the model, initialize parameters, set the maximum number of iterations, determine the values ​​of the inertia weight w, learning factors c1 and c2, and the parameters of NSGA;

[0009] S3: Optimize the expression of each particle’s solution to FJSP through dual encoding, using two-level encoding for each chromosome and combining them together;

[0010] S4: Roulette initialization is adopted, with the maximum completion time of machine tools and workshop energy consumption as fitness functions, and the optimization target weights are 0.6 and 0.4 respectively, and then the initialization population is generated;

[0011] S5: Merge the particle groups into a population, perform non-dominated sorting on the population, and divide different Pareto frontier layers; then, calculate the crowding degree of the solution in each frontier layer, and perform selection operations based on the non-dominated sorting results and crowding degree to select excellent individuals, perform crossover processing on the selected individuals, generate the next generation of individuals, and then perform mutation processing on the next generation of individuals to improve the diversity of the population;

[0012] S6: According to the particle motion rules, the particle speed and position are updated, and the fitness of the updated position is calculated. If the fitness of the new position is better than the fitness of the individual optimal solution, the individual optimal solution is updated to find the global optimal solution.

[0013] S7: Repeat steps S5 and S6 until the fitness reaches an acceptable range or reaches the set maximum number of iterations, and finally output the final Pareto frontier solution set, that is, the optimization result.

[0014] In this paper, a hybrid algorithm of MOPSO and NSGA-II is used to optimize and solve the flexible job shop scheduling problem considering machine pre-maintenance and shutdown / restart strategies; secondly, by analyzing the impact of machine tool pre-maintenance and shutdown / restart strategies on the energy consumption of flexible job shops, a flexible job shop scheduling problem model considering energy consumption is established; thirdly, the MOPSO algorithm is used to perform operations such as population initialization and update of the NSGA-II algorithm to achieve algorithm optimization; finally, the effectiveness of the model and method is verified through examples.

[0015] Furthermore, in step S1, the setting of the established model includes: the initial engraving workshop has a total of n types of workpieces to be processed n i 、m available machine tools M m , where workpiece n i The batch size is q i , each workpiece n i Including s process o ij; When the machine tool is performing pre-maintenance, shutdown and restart strategies, the machine tool is in an inactive state.

[0016] Furthermore, in step S1, the objective function includes:

[0017] maxC m m∈[1,k];

[0018] Among them, C m is the completion time, which is the maximum value of the completion time of all processes of all workpieces processed by all machines;

[0019]

[0020] Among them, TE is the total energy consumption of the machine tool in the flexible job shop processing process, AE is the energy consumption of idle cutting, CE is the energy consumption of the machine tool during the cutting processing period, and IE is the idle energy consumption of the machine tool.

[0021] Further, in step S1, according to the workpieces to be processed of the flexible job shop production line and the shop configuration information table, it is determined that a constraint model is established based on the flexible job shop scheduling problem, the process sequence, and whether the machine tool is available, including:

[0022] A1: Energy consumption constraint model;

[0023]

[0024] Where P au is the auxiliary system power, P u is the no-load power of the machine tool, The tool setting time required before machine tool processing;

[0025]

[0026] Where P cut is the cutting power, For process o ijr In machine tool M m Upper machining cutting time;

[0027]

[0028] Where P st is the machine tool standby power, Indicates machine tool M m Idle waiting time, y m is a 0-1 decision variable, machine tool M m Implement shutdown / restart strategy, y m The value is 1, otherwise it is 0;

[0029] A2: Production constraints;

[0030] Ensure that each process can only be processed by one machine tool, and the workpiece processing needs to meet the process processing sequence constraints of the same workpiece:

[0031]

[0032] and

[0033] In the formula, x ijm is a 0-1 variable, x ijmo =1 means that in the jth process route, process o ij Select Machine Tool M m ;

[0034] At the same time, a machine tool can only process one process of a workpiece:

[0035]

[0036] In the formula, Indicates machine tool M m The start time of the rth process, Indicates machine tool M m The completion time of the rth process;

[0037] A3: Pre-maintenance constraints:

[0038] Machine tool pre-maintenance tasks and process processing cannot be performed at the same time:

[0039]

[0040]

[0041] In the formula, For machine tool M m The start time of the scheduled maintenance, For machine tool M m The time when the scheduled maintenance ends;

[0042] At any time t, the machine tool equipment M involved in the processing m Its reliability R m It cannot be lower than the reliability threshold R0, otherwise the machine tool will be required to perform pre-maintenance:

[0043] R m >R0;

[0044]

[0045] Where R mn (t) represents the machine tool M m Reliability function at time t;

[0046] A4: Machine shutdown and restart constraints;

[0047] The shortest time required to implement the machine shutdown and restart strategy is not less than the machine's no-load balance time:

[0048]

[0049] In the formula, For machine tool M m The completion time of the rth processing step.

[0050] Furthermore, in step S2, the parameters of NSGA include NSGA-II crossover probability, mutation probability, tournament selection scale parameter, and maximum number of iterations; based on the experimental results, the impact of the parameters on the algorithm performance is analyzed, and the parameters are gradually adjusted to find the optimal or better parameter combination.

[0051] Further, in step S3, the dual encoding method includes: using encoding based on process ordering, each gene in this layer represents the sequential processing relationship between different processing processes of all workpieces to be processed; encoding based on machine allocation vector, each gene in this layer of encoding represents the corresponding selected processing machine, wherein each process corresponds to multiple processing machines in the FJSP problem;

[0052] In the solution space, each particle contains two gene segments, MA and OS, which are spliced ​​together to form a particle with a length of 2T0 during the encoding process. Represents the sum of all process numbers of the workpiece, and the particles in the solution space are expressed in the form of MA+OS.

[0053] Preferably, in step S5, the population iteration method includes the following steps:

[0054] B1: Initialize the roulette wheel selection P t The populations are merged, and after non-dominated sorting, selection, crossover and mutation, the offspring population Q is generated. t , merge the two populations to form a population R of size 2N t ;

[0055] B2: The population R of size 2N t Perform fast non-dominated sorting and calculate the crowding degree of individuals in each non-dominated layer. Select appropriate individuals to form a new parent population P according to the non-dominated relationship and the crowding degree of the individuals. t+1 ;

[0056] B3: Generate a new offspring population Q through NSGA-Ⅱ t+1 , P t+1 With Q t+1 Merge to form a new population Rt , repeat the above steps until the maximum number of iterations is reached.

[0057] Optimally, in step S6, the particle update model is:

[0058] v ij (t+1)=w×v ij (t)+c1r1[p ij (t)-x ij (t)]+c2r2[p gj (t)-x ij (t)];

[0059] x ij (t+1)=x ij (t)+v ij (t+1);

[0060] Where w represents the extent to which the particle retains its velocity.

[0061] When the value of w is large, the method has a stronger global optimization ability; when the value of w is small, the algorithm has a better local optimization performance. Usually, the inertia weight decreases as the number of iterations increases.

[0062] Optimally, in step S7, the fitness is to set the thresholds of the maximum completion time and total energy consumption. If the fitness is within an acceptable range, a better solution that meets the requirements has been found and the iteration can be stopped; the maximum number of iterations is the maximum number set to prevent the algorithm from falling into an infinite loop.

[0063] In each iteration, the particle speed and position are first updated according to the particle motion rules, and then the fitness of the updated position is calculated. If the fitness of the new position is better than the fitness of the individual optimal solution, the individual optimal solution is updated. Next, determine whether the fitness is within an acceptable range. The fitness is based on the thresholds of the maximum completion time and total energy consumption. If the fitness is within an acceptable range, it means that a better solution that meets the requirements has been found and the iteration can be stopped. In addition, determine whether the set maximum number of iterations has been reached. If the maximum number of iterations has been reached, the iteration is stopped to prevent the algorithm from falling into an infinite loop.

[0064] Finally, the final Pareto frontier solution set is output. The Pareto frontier solution set is a collection of non-dominated solutions found in multiple iterations. These solutions weigh and optimize between the maximum completion time and total energy consumption, representing different optimal solutions. By outputting the Pareto frontier solution set, the most appropriate scheduling solution can be selected according to actual needs, thereby achieving the optimization goal of flexible job shop scheduling.

[0065] The present invention aims at the problems existing in flexible job shop scheduling, such as machine tool equipment failure leading to processing interruption, increased energy consumption, increased total energy consumption of the workshop, and machine idle waiting. The present invention is mainly divided into three parts: a flexible job shop scheduling problem model is established, taking into account factors such as machine tool pre-maintenance, shutdown / restart strategy and energy consumption constraints; the MOPSO and NSGA-II hybrid algorithm combines the fast search and preliminary optimization capabilities of the particle swarm algorithm with the crossover and mutation operations of the non-dominated sorting genetic algorithm to increase the diversity of solutions; double-layer encoding, roulette population initialization and particle update operations help to improve the algorithm's solution efficiency and quality; and example verification ensures the effectiveness of the model and method in practical applications.

[0066] Beneficial effects: Compared with the prior art, the advantages of the present invention are:

[0067] The present invention can effectively solve the problems of extended processing time and increased total energy consumption of the workshop caused by equipment failure and idle machine tools in workshop production scheduling. By considering machine tool pre-maintenance, shutdown / restart strategy and energy consumption constraints, a flexible workshop energy consumption model is established, and the MOPSO and NSGA-II hybrid algorithm is used for solving it, which can take into account the solution speed and global search capability to a certain extent, effectively reduce energy consumption and improve production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 It is a schematic diagram of the framework of the present invention;

[0069] Figure 2 It is the overall flow chart of the MOPSO and NSGA-II hybrid algorithm of the present invention;

[0070] Figure 3 The present invention does not consider the initial scheduling scheme of each strategy;

[0071] Figure 4 The present invention considers three different strategy combination schemes for the scheduling scheme Gantt chart;

[0072] Figure 5 It is the energy consumption composition analysis diagram of each scheme of the present invention;

[0073] Figure 6 It is a time composition analysis diagram of each scheme of the present invention. DETAILED DESCRIPTION

[0074] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention.

[0075] A flexible job shop scheduling optimization method considering machine predictive maintenance and controllable working hours, such as Figures 1 to 6 As shown, the following steps are included:

[0076] S1: Analyze the impact of machine tool pre-maintenance and shutdown / restart strategy on the energy consumption of the Flexible Job-shop Scheduling Problem (FJSP), and model the FJSP considering machine tool pre-maintenance and shutdown / restart strategy, including objective function and constraint function.

[0077] S2: The MOPSO and NSGA-II hybrid algorithm is used to optimize and solve the model, initialize the parameters, set the maximum number of iterations, determine the values ​​of the inertia weight w, learning factors c1 and c2, and related parameters of the non-dominated genetic algorithm.

[0078] S3: A dual encoding method is used to enable each particle to better express the solution of FJSP. Two-level encoding is used for each chromosome and they are combined together.

[0079] S4: Using the roulette initialization method, the maximum completion time of the machine tool and the energy consumption of the workshop are used as the fitness function, the optimization target weights are 0.6 and 0.4 respectively, and then the initialization population is generated.

[0080] S5: Merge the particle groups into a population, perform non-dominated sorting on the population, and divide different frontier layers (Pareto frontiers); then, calculate the crowding degree of the solution in each frontier layer, and perform selection operations based on the non-dominated sorting results and crowding degree to select excellent individuals, perform crossover processing on the selected individuals, generate the next generation of individuals, and then perform mutation processing on the next generation of individuals to improve the diversity of the population.

[0081] S6: Find the global optimal solution. Update the particle speed and position according to the particle motion rules. Calculate the fitness of the updated position. If the fitness of the new position is better than the fitness of the individual optimal solution, update the individual optimal solution.

[0082] S7: Repeat steps S5 and S6 until the fitness reaches an acceptable range or reaches the set maximum number of iterations. Finally, the final Pareto frontier solution set is output, that is, the optimization result.

[0083] Based on the above scheme, in this embodiment, the above flexible job shop scheduling optimization method is applied as an example, as follows:

[0084] The experimental data of this example comes from a flexible operation workshop with 8 machine tools and 3 types of workpieces to be processed. The machine tool power information is shown in Table 2, the processing batch, arrival time, and delivery date are shown in Table 1, and the machine tool information that can be selected under different process routes of each workpiece in the workshop is shown in Table 3.

[0085] Table 1 Workpiece processing information

[0086]

[0087] Table 2 Machine tool power information

[0088]

[0089] Table 3 Workpiece processing procedures and optional machine tool information

[0090]

[0091] In the experiment, a case analysis was conducted on a flexible workshop. The workshop has 8 machine tools and 3 types of workpieces to be processed. The experiment was conducted by setting relevant information. During the experiment, the machine tool power information, processing batch, arrival time, delivery date, and the machine tool information that can be selected under different process routes of each workpiece were obtained.

[0092] The overall technical framework is as follows Figure 1 As shown in the figure, the specific scheduling process is:

[0093] 1) The assumptions of FJSP modeling considering machine preventive maintenance and shutdown / restart strategy are:

[0094] There are n types of workpieces to be processed in the initial engraving workshop. i 、m available machine tools M m , where workpiece n i The batch size is q i , each workpiece n i The process includes o ij When the machine is undergoing a preventive maintenance plan and shutdown / restart strategy, the machine is in an out-of-service state.

[0095] 2) The FJSP optimization objective function considering machine pre-maintenance and shutdown / restart strategy is:

[0096] maxC m m∈[1,k]

[0097] Where completion time C m It is the maximum value of the completion time of all processes of all workpieces processed by all machine tools.

[0098]

[0099] In the formula, the total energy consumption TE of machine tools in the machining process of flexible workshop is mainly composed of five parts, including the energy consumption of workpiece clamping and disassembly, the energy consumption of idle cutting AE, the energy consumption of machine tools during cutting processing CE, the energy consumption of tool changing, and the energy consumption of machine tools idle IE. Among them, the energy consumption of workpiece clamping and disassembly and the energy consumption of tool changing have little effect on the total energy consumption of the workshop and are not considered.

[0100] 3) The constraint model is:

[0101] According to the processing workpieces of the flexible job shop production line and the shop configuration information table, the constraints based on the flexible job shop scheduling problem and the process sequence and whether the machine is available are determined.

[0102] A1: Energy consumption constraint model;

[0103]

[0104] Where P au is the auxiliary system power, P u is the no-load power of the machine tool, The tool setting time required before machine tool processing;

[0105]

[0106] Where P cut is the cutting power, For process o ijr In machine tool M m Upper machining cutting time;

[0107]

[0108] Where P st is the machine tool standby power, Indicates machine tool M m Idle waiting time, y m is a 0-1 decision variable, machine tool M m Implement shutdown / restart strategy, y m The value is 1, otherwise it is 0;

[0109] A2: Production constraints;

[0110] Ensure that each process can only be processed by one machine, and the workpiece processing needs to meet the process processing sequence constraints of the same workpiece:

[0111]

[0112] and

[0113] Where x ijm is a 0-1 variable, x ijmo =1 means that in the jth process route, process o ij Select Machine Tool M m .

[0114] At the same time, a machine tool can only process one process of a workpiece:

[0115]

[0116] In the formula Indicates machine tool M m The start time of the rth process, Indicates machine tool M m The completion time of the rth processing step.

[0117] A3: Pre-maintenance constraints:

[0118] Machine tool pre-maintenance tasks and process processing cannot be performed at the same time:

[0119]

[0120] In the formula, For machine tool M m The start time of the scheduled maintenance, For machine tool M m The time when the scheduled maintenance ends;

[0121] At any time t, the machine tool equipment M involved in the processing m Its reliability R m It cannot be lower than the reliability threshold R0, otherwise the machine tool will be required to perform pre-maintenance:

[0122] R m >R0

[0123]

[0124] Where R mn (t) represents the machine tool M m Reliability function at time t;

[0125] A4: Machine shutdown / restart constraints

[0126] The shortest time required to implement the machine shutdown / restart strategy is not less than the machine's no-load balance time:

[0127]

[0128] In the formula For machine tool M m The completion time of the rth processing step.

[0129] 4) The algorithm parameters are set as:

[0130] Determine the relevant parameters of MOPSO such as inertia weight w, learning factors c1 and c2, as well as relevant parameters of NSGA-II such as crossover probability, mutation probability and tournament selection scale, and finally determine the maximum number of iterations.

[0131] According to the experimental results, analyze the impact of parameters on algorithm performance, gradually adjust the parameters, and find the optimal or better parameter combination.

[0132] 5) The double-layer encoding method is:

[0133] Operation Sequence (OS): Each gene in this layer represents the order of processing between different processing steps of all workpieces to be processed. Machine Assignment (MA): In the FJSP problem, each process corresponds to multiple processing machines, and each gene in this layer of coding represents the corresponding processing machine selected.

[0134] In the solution space, each particle contains two gene segments, MA and OS. During the encoding process, they are spliced ​​together to form a particle with a length of 2T0, where It represents the sum of all process numbers of the workpiece, so the particles in the solution space can be expressed in the form of MA+OS.

[0135] 6) The method of population initialization is:

[0136] When solving the FJSP problem, for the intelligent optimization algorithm, generating an initial population is the starting point of the algorithm's optimization. Therefore, its quality, in a sense, has a great impact on the quality of the final optimal solution.

[0137] A high-quality initial population can improve the convergence speed of the algorithm. Considering that the optimization goal is to minimize the maximum completion time and total energy consumption, the roulette wheel selection population initialization method is used to generate the initial population. In this method, the selection strategy increases the chances of excellent individuals being selected in the algorithm, which helps the algorithm converge to the optimal solution faster.

[0138] The roulette wheel selection method for population initialization was used, with the maximum completion time of machine tools and the energy consumption of the workshop as the fitness function, and the optimization target weights as 0.6 and 0.4, respectively, to generate the initial population.

[0139] 7) The method of population iteration is:

[0140] B1: Initialize the roulette wheel selection P t The populations are merged, and after non-dominated sorting, selection, crossover and mutation, the offspring population Q is generated. t , merge the two populations to form a population R of size 2N t ;

[0141] B2: The population R of size 2N t Perform fast non-dominated sorting and calculate the crowding degree of individuals in each non-dominated layer. Select appropriate individuals to form a new parent population P according to the non-dominated relationship and the crowding degree of the individuals. t+1 ;

[0142] B3: Generate a new offspring population Q through the basic operations of NSGA-Ⅱ t+1 , P t+1 With Q t+1 Merge to form a new population R t , repeat the above steps until the maximum number of iterations is reached.

[0143] 8) The particles are updated as follows:

[0144] v ij (t+1)=w×v ij (t)+c1r1[p ij (t)-x ij (t)]+c2r2[p gj (t)-x ij (t)]

[0145] x ij (t+1)=x ij (t)+v ij (t+1)

[0146] Where w represents the extent to which the particle retains the particle's velocity; when the value of is large, the method has a stronger global optimization ability; when the value of w is small, the algorithm has better local optimization performance. Usually, the inertia weight is reduced as the number of iterations increases.

[0147] 9) Iterate and output the Pareto frontier solution set:

[0148] In the process of solving the problem, steps S5 and S6 need to be repeated to continuously iterate and update the position and speed of the particles to find the optimal solution. Specifically, in each iteration, the particle speed and position are first updated according to the particle motion rules, and then the fitness of the updated position is calculated. If the fitness of the new position is better than the fitness of the individual optimal solution, the individual optimal solution is updated. Next, determine whether the fitness is within an acceptable range. The fitness is based on the thresholds of the maximum completion time and total energy consumption. If the fitness is within an acceptable range, it means that a better solution that meets the requirements has been found and the iteration can be stopped. In addition, determine whether the set maximum number of iterations has been reached. If the maximum number of iterations has been reached, stop the iteration to prevent the algorithm from falling into an infinite loop.

[0149] Finally, the final Pareto frontier solution set is output. The Pareto frontier solution set is a collection of non-dominated solutions found in multiple iterations. These solutions weigh and optimize between the maximum completion time and total energy consumption, representing different optimal solutions. By outputting the Pareto frontier solution set, the most appropriate scheduling solution can be selected according to actual needs, thereby achieving the optimization goal of flexible job shop scheduling.

[0150] In order to verify the effect of the flexible job shop scheduling method proposed in this invention, the experimental data of this example comes from a flexible job shop with 8 machine tools and 3 types of workpieces to be processed. According to the above machine tool and workpiece information, the initial scheduling plan is generated based on the hybrid algorithm of MOPSO and NSGA-Ⅱ with the maximum completion time and the total energy consumption of the workshop as the goal, as shown in the following figure: Figure 3 In addition, the scheduling schemes considering the periodic maintenance strategy, the periodic maintenance strategy and the shutdown / restart strategy, and the scheduling schemes considering the machine tool pre-maintenance and the machine tool shutdown / restart strategy are shown as follows: Figure 4 As shown in (a), (b), and (c), it can be seen from the comparison of the figures that the method proposed in the present invention re-arranges the process flow reasonably under the consideration of the preventive maintenance strategy and arranges the shutdown of the machine tool with a large amount of idle waiting time considering the machine tool shutdown / restart strategy; in order to analyze the energy consumption and time composition, the energy consumption data of each scheme and the time composition are analyzed and compared. The comparison results are shown in FIG. Figure 5 , Figure 6 As can be seen from the figure, cutting energy consumption accounts for the largest proportion, followed by machine idle energy consumption and empty cutting energy consumption; the time composition is different, cutting time accounts for the largest proportion, followed by idle time, and empty cutting time accounts for the smallest proportion. Comparison of different schemes shows that the scheduling scheme considering machine tool pre-maintenance and machine tool shutdown / restart strategy reduces total energy consumption by 5.8% and shortens the maximum completion time by 4.4%.

Claims

1. A flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours, characterized by The following steps are involved: S1: Analyze the impact of machine tool pre-maintenance, machine tool shutdown and restart strategies on the energy consumption of the flexible job shop, and model the FJSP considering machine tool pre-maintenance, machine tool shutdown and restart strategies. The model includes objective function and constraint model. S2: Use the MOPSO and NSGA-II hybrid algorithm to optimize and solve the model, initialize the parameters, set the maximum number of iterations, determine the values ​​of the inertia weight w, learning factors c1 and c2, and the parameters of NSGA; S3: Optimize the expression of each particle's solution to FJSP through dual encoding, using two-level encoding for each chromosome and combining them together; S4: Roulette initialization is adopted, with the maximum completion time of machine tools and the energy consumption of workshops as the fitness function, and the optimization target weights are 0.6 and 0.4 respectively, and then the initialization population is generated; S5: Merge the particle groups into a population, perform non-dominated sorting on the population, and divide different Pareto frontier layers; then, calculate the crowding degree of the solution in each frontier layer, and perform selection operations based on the non-dominated sorting results and crowding degree to select excellent individuals, perform crossover processing on the selected individuals, generate the next generation of individuals, and then perform mutation processing on the next generation of individuals to improve the diversity of the population; S6: According to the particle motion rules, the particle speed and position are updated, and the fitness of the updated position is calculated. If the fitness of the new position is better than the fitness of the individual optimal solution, the individual optimal solution is updated to find the global optimal solution. S7: Repeat steps S5 and S6 until the fitness reaches an acceptable range or reaches the set maximum number of iterations, and finally output the final Pareto frontier solution set, that is, the optimization result.

2. The flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours according to claim 1 is characterized in that: In step S1, the settings of the established model include: the initial engraving workshop has a total of n types of workpieces to be processed i 、m available machine tools M m , where workpiece n i The batch size is q i , each workpiece n i Including s process o ij ; When the machine tool is performing pre-maintenance, shutdown and restart strategies, the machine tool is in an inactive state.

3. The flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours according to claim 1 is characterized in that: In step S1, the objective function includes: maxC m m∈[1,k]; Among them, C m is the completion time, which is the maximum value of the completion time of all processes of all workpieces processed by all machines; Among them, TE is the total energy consumption of the machine tool in the flexible job shop processing process, AE is the energy consumption of idle cutting, CE is the energy consumption of the machine tool during the cutting processing period, and IE is the idle energy consumption of the machine tool.

4. The flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours according to claim 1 is characterized in that: In step S1, according to the workpieces to be processed of the flexible job shop production line and the shop configuration information table, it is determined that a constraint model is established based on the flexible job shop scheduling problem, the process sequence, and whether the machine tool is available, including: A1: Energy consumption constraint model; Where P au is the auxiliary system power, P u is the no-load power of the machine tool, The tool setting time required before machine tool processing; Where P cut is the cutting power, For process o ijr In machine tool M m Upper machining cutting time; Where P st is the machine tool standby power, Indicates machine tool M m Idle waiting time, y m is a 0-1 decision variable, machine tool M m Implement shutdown / restart strategy, y m The value is 1, otherwise it is 0; A2: Production constraints; Ensure that each process can only be processed by one machine tool, and the workpiece processing needs to meet the process processing sequence constraints of the same workpiece: and In the formula, x ijm is a 0-1 variable, x ijmo =1 means that in the jth process route, process o ij Select Machine Tool M m ; At the same time, a machine tool can only process one process of a workpiece: In the formula, Indicates machine tool M m The start time of the rth process, Indicates machine tool M m The completion time of the rth process; A3: Pre-maintenance constraints: Machine tool pre-maintenance tasks and process processing cannot be performed at the same time: In the formula, For machine tool M m The start time of the scheduled maintenance, For machine tool M m The time when the scheduled maintenance ends; At any time t, the machine tool equipment M involved in the processing m Its reliability R m It cannot be lower than the reliability threshold R0, otherwise the machine tool will be required to perform pre-maintenance: R m >R0; Where R mn (t) represents the machine tool M m Reliability function at time t; A4: Machine shutdown and restart constraints; The shortest time required to implement the machine shutdown and restart strategy is not less than the machine's no-load balance time: In the formula, For machine tool M m The completion time of the rth processing step.

5. The flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours according to claim 1 is characterized in that: In step S2, the parameters of NSGA include NSGA-II crossover probability, mutation probability, tournament selection scale parameter, and maximum number of iterations. Based on the experimental results, the impact of the parameters on the algorithm performance is analyzed, and the parameters are gradually adjusted to find the optimal or better parameter combination.

6. The flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours according to claim 1 is characterized in that: In step S3, the dual encoding method includes: using encoding based on process ordering, each gene in this layer represents the sequential processing relationship between different processing processes of all workpieces to be processed; encoding based on machine allocation vector, each gene in this layer of encoding represents the corresponding selected processing machine, wherein each process in the FJSP problem corresponds to multiple processing machines; In the solution space, each particle contains two gene segments, MA and OS, which are spliced ​​together to form a particle with a length of 2T0 during the encoding process. Represents the sum of all process numbers of the workpiece, and the particles in the solution space are expressed in the form of MA+OS.

7. The flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours according to claim 1 is characterized in that: In step S5, the population iteration method includes the following steps: B1: Initialize the roulette wheel selection P t The populations are merged, and after non-dominated sorting, selection, crossover and mutation, the offspring population Q is generated. t , merge the two populations to form a population R of size 2N t ; B2: The population R of size 2N t Perform fast non-dominated sorting and calculate the crowding degree of individuals in each non-dominated layer. Select appropriate individuals to form a new parent population P according to the non-dominated relationship and the crowding degree of the individuals. t+1 ; B3: Generate a new offspring population Q through NSGA-Ⅱ t+1 , P t+1 With Q t+1 Merge to form a new population R t , repeat the above steps until the maximum number of iterations is reached.

8. The flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours according to claim 1 is characterized in that: In step S6, the model of particle update is: v ij (t+1)=w×v ij (t)+c1r1[p ij (t)-x ij (t)]+c2r2[p gj (t)-x ij (t)]; x ij (t+1)=x ij (t)+v ij (t+1); Where w represents the extent to which the particle retains its velocity.

9. The flexible job shop scheduling optimization method considering machine pre-maintenance and controllable working hours according to claim 1 is characterized in that: In step S7, the fitness is to set the thresholds of the maximum completion time and the total energy consumption. If the fitness reaches an acceptable range, a better solution that meets the requirements has been found and the iteration can be stopped; the maximum number of iterations is the maximum number set to prevent the algorithm from falling into an infinite loop.

Images