A simulation-based optimization method for production planning of tank welding workshop

Abstract

The present invention discloses a method for optimizing the production plan of a tank welding workshop based on simulation. The simulation model of the tank welding workshop is constructed to accurately describe the tank welding workshop, and a heuristic optimization method is used in combination with simulation to continuously iterate and find the best production plan. The early completion time of this plan is controllable, and there is no delayed completion phenomenon. By optimizing the continuous production time arrangement, the problem of early / delayed completion of the workshop can be effectively solved, and the cost problem caused by early / delayed completion of the workshop can be reduced, and the delay / early completion problem caused by relying on experience to formulate the production plan in the traditional method can be solved. At the same time, the production plan obtained by simulation is more real and reliable, and can provide active guidance for the production plan of the tank welding workshop.

Description

Technical Field

[0001] The invention belongs to the field of production control, and more specifically, relates to a simulation-based method for optimizing the production plan of a tank welding workshop. Background Art

[0002] Manufacturing plants are becoming more and more complex, which is reflected in many aspects such as factory facilities, logistics strategies, product portfolios, and abnormal factors. Intelligent design is inseparable from scientific simulation analysis. Based on the factory simulation model, low-cost, fast, and scientific factory design can be achieved through simulation, verification, and optimization. Furthermore, when products, processes, or logistics change, simulation analysis and optimization of production lines can help to continuously ensure factory efficiency. Product delays will reduce the company's service level and affect cooperation between companies; at the same time, it will affect the production process of downstream products, resulting in huge delay costs; early completion of products will increase workshop inventory, logistics management, and labor costs. Solving the problem of delayed / early completion planning optimization is of great significance to enterprises.

[0003] The optimization of the production plan of a tank welding workshop based on simulation mainly solves the optimization problem of the production time and production sequence of products in a complex workshop with multiple varieties of products. In the traditional planning and arrangement of a tank welding workshop, the formulation of the plan is generally arranged by relevant staff based on experience, lacking consideration of the overall plan and scientific quantitative analysis, resulting in a large difference between the actual production process and the plan, and problems such as long early completion time and delayed completion. With the development of computing and operations research, mathematical models have been introduced for various types of workshops to solve them through intelligent optimization methods. However, for complex workshops with logistics, human resource scheduling and random factors, mathematical models that are simple and abstract and ignore detail factors often cannot accurately describe complex workshops, and have little effect on optimizing the production plan of complex workshops. Summary of the invention

[0004] In view of the above defects or improvement needs of the prior art, the present invention provides a simulation-based method for optimizing the production plan of a tank welding workshop. The method uses a heuristic optimization method combined with simulation iteration to find the best production plan, optimizes the continuous production schedule, and effectively solves the problem of early / delayed completion of the workshop, thereby reducing the cost problem caused by early / delayed completion of the workshop.

[0005] To achieve the above object, according to a first aspect of the present invention, a method for optimizing the production plan of a tank welding workshop based on simulation is provided, comprising:

[0006] S1, establish a tank welding workshop planning optimization simulation model, input the production plan and planned completion time into the simulation model, and obtain the actual completion time of each product;

[0007] S2. Arrange all products in ascending order according to the planned completion time, and let the iteration number k = 0;

[0008] S3. Let i = the total number of products;

[0009] S4. Determine whether the i-th product is overdue. If so, go to step S5; if not, determine whether the i-th product is completed ahead of schedule. If so, go to step S6; if not, go to step S8;

[0010] S5. Let the actual start time s i = s i - t 3 , and go to step S7; where t 3 = μ × t 1 , t 1 is the overdue time, and μ is the overdue adjustment factor;

[0011] S6. If s i + t 4 > s i+1 - T 2 , then s i = s i+1 - T 2 ; otherwise, s i = s i + t 4 , and go to step S7; where t 4 = λ × t 2 , t 2 is the early time, λ is the early adjustment factor, and T 2 is the minimum start interval;

[0012] S7. Input the updated production plan into the simulation model to obtain the actual completion time of each product. If T 1 < C, end; otherwise, k = k + 1; if k%A = 0, then T 1 = T 1 - B, and return to step S3; otherwise, directly return to step S3; where T 1 is the allowable early completion time;

[0013] S8. Let i = i - 1. If i > 0, return to step S4; otherwise, end.

[0014] Preferably, if the overdue time t 1 > t 1 `, then the i-th product is overdue; where t 1 ` is the allowable overdue time, and t 1 `≥0.

[0015] Preferably, it is characterized in that 0<λ<1, T 1 ≥0.

[0016] Preferably, the production plan includes the planned production start time of each product, the production sequence of each part of each product, and the production start time of each part of each product.

[0017] Preferably, before step S1, the method further includes: optimizing the production sequence of each product and each part using an intelligent optimization algorithm.

[0018] Preferably, the intelligent optimization algorithm includes: GA algorithm, GA algorithm improved based on migratory bird optimization algorithm, particle swarm algorithm, taboo search algorithm, migratory bird optimization algorithm.

[0019] Preferably, when the intelligent optimization algorithm is a GA algorithm improved based on the migratory bird optimization algorithm, the optimization steps include:

[0020] A1, define the fitness function and set the number of iterations, randomly initialize the coding sequence of individuals in the population, and divide the population structure into leading birds and left and right queues;

[0021] The coding sequence is a decimal integer sequence containing n elements, where n is the number of parts of the product; the value of the element is the number of the part type, and the position of the element in the coding sequence is the production order of the corresponding parts;

[0022] A2, after performing crossover and mutation operations on the codes of each individual, a neighborhood search is performed according to the neighborhood structure, and multiple new offspring individuals are obtained by combining the sharing mechanism of other individuals;

[0023] A3, decode each individual to obtain the production sequence of product parts, and input it into the simulation model to calculate its fitness; where the fitness function is: f = min (1 / T), T is the total completion time;

[0024] A4, sort the fitness values ​​of individuals based on the competition mechanism. After all individuals are sorted, they are divided into leading birds, left and right queues, and a new population is formed;

[0025] A5, determine whether the number of iterations has been reached, if so, output the optimal production sequence result, if not, return to step A2.

[0026] Preferably, a partial mapping crossover operator is used for the crossover operation, and an exchange mutation operator is used for the mutation operation;

[0027] The neighborhood structure adopts selective insertion and greedy insertion.

[0028] Preferably, the method for establishing the tank welding workshop planning optimization simulation model is:

[0029] B1, import the equipment layout, process technology, production plan, workshop equipment and resource list of the tank welding workshop into the layout CAD model;

[0030] B2, create model objects such as equipment, buffer, logistics cart, data table, etc. in sequence; define logistics scheduling module and worker scheduling module;

[0031] B3 edits the logic script in the simulation model to ensure that the simulation runs according to the actual workshop process route and process time;

[0032] B4, define the production plan verification module: collect the production time and production sequence plan data of the products produced in the workshop; input the data into the simulation, decompose it into product parts list according to the production sequence, and set the simulation model production plan according to the production time; set the simulation model start date and work calendar; run the simulation, obtain the actual completion time of each product, and compare it with the planned completion time.

[0033] According to a second aspect of the present invention, there is provided a simulation-based tank welding shop production planning optimization system, characterized in that it comprises: a computer-readable storage medium and a processor;

[0034] The computer-readable storage medium is used to store executable instructions;

[0035] The processor is used to read the executable instructions stored in the computer-readable storage medium and execute the method as described in the first aspect.

[0036] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0037] 1. Based on the characteristics of the tank welding workshop, the present invention provides a method for optimizing the production plan of the tank welding workshop based on simulation. The simulation model of the tank welding workshop is constructed to accurately describe the tank welding workshop, and the heuristic optimization method is used to dynamically adjust the production plan of all products in the workshop in combination with the simulation model, so as to obtain the optimized production plan of the tank welding workshop. The early completion time of this plan is controllable, and there is no delayed completion phenomenon. By optimizing the continuous production time arrangement, the problem of early / delayed completion of the workshop can be effectively solved, and the cost problem caused by early / delayed completion of the workshop can be reduced; the problem of delayed / early completion caused by relying on experience to formulate the production plan in the traditional method is solved. At the same time, the production plan obtained by simulation is more real and reliable, and can provide active guidance for the production plan of the tank welding workshop.

[0038] 2. The simulation-based tank welding workshop production plan optimization method provided by the present invention optimizes the production sequence of different products in the tank welding workshop based on an intelligent optimization algorithm, such as a GA algorithm improved based on a migratory bird algorithm, before dynamically adjusting and optimizing the production plan. This can reduce the maximum completion time of the product and save workshop production costs, thereby further improving the optimization effect.

[0039] 3. The simulation-based tank welding workshop production plan optimization method provided by the present invention constructs a tank welding workshop simulation model through information such as workshop equipment layout, production plan, logistics method, process flow, etc., and intuitively reflects the complexity and randomness of the workshop by constructing a simulation model, thus overcoming the problem that traditional planning optimization methods cannot accurately describe the workshop model.

[0040] 4. The simulation-based tank welding workshop production plan optimization method provided by the present invention is simple and easy to implement, fast and highly adaptable, and can flexibly respond to the planned optimization arrangements after the workshop product production is adjusted. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 One of the flow diagrams of the method for optimizing the production plan of a tank welding workshop based on simulation provided by an embodiment of the present invention;

[0042] Figure 2 A GA flow chart based on the improved migratory bird optimization algorithm provided in an embodiment of the present invention;

[0043] Figure 3 A second flow chart of a method for optimizing a tank welding workshop production plan based on simulation provided by an embodiment of the present invention;

[0044] Figure 4 A workshop layout diagram provided by an embodiment of the present invention;

[0045] Figure 5 Product process route and production data diagram provided for the embodiment of the present invention;

[0046] Figure 6 BOM diagram of products produced by a tank welding workshop provided in an embodiment of the present invention;

[0047] Figure 7 A schematic diagram of production plan verification provided by an embodiment of the present invention;

[0048] Figure 8 This is a comparison chart between the planned and actual completion time before optimization;

[0049] Fig. 9 This is a comparison chart between the optimized plan and the actual completion time. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0051] The embodiment of the present invention provides a method for optimizing the production plan of a tank welding workshop based on simulation, such as Figure 1 As shown, including:

[0052] S1, establish a tank welding workshop planning optimization simulation model, input the original production plan and planned completion time into the simulation model, run the simulation, and obtain the actual completion time of each product.

[0053] Specifically, the production plan includes the planned production time of each product, the production sequence of each part of each product, and the production time of each part of each product.

[0054] The tank welding workshop production is simulated by the tank welding workshop planning optimization simulation model, the actual completion time of each product under the simulation is obtained, and the actual completion time is compared with the actual completion time to achieve dynamic adjustment of the production plan.

[0055] Preferably, the method for establishing the tank welding workshop planning optimization simulation model is:

[0056] B1, import the equipment layout, process technology, production plan, workshop equipment and resource list of the tank welding workshop into the layout CAD model;

[0057] B2, create model objects such as equipment, buffer, logistics cart, data table, etc. in sequence; define logistics scheduling module and worker scheduling module;

[0058] B3 edits the logic script in the simulation model to ensure that the simulation runs according to the actual workshop process route and process time;

[0059] B4, define the production plan verification module: collect the production time and production sequence plan data of the products produced in the workshop; input the data into the simulation, decompose it into product parts list according to the production sequence, and set the simulation model production plan according to the production time; set the simulation model start date and work calendar; run the simulation, obtain the actual completion time of each product, and compare it with the planned completion time.

[0060] Specifically, the information of workshop equipment layout, production plan, logistics mode, process flow and so on is obtained to build a simulation model of tank welding workshop. The specific steps are as follows:

[0061] Step 1. Systematic research, the research content mainly includes the equipment layout, process technology, production plan, workshop equipment, resource list, etc. of the tank welding workshop; the simulation software FactorySimulation independently developed by Huazhong University of Science and Technology is used for modeling;

[0062] Step 2. Based on the information obtained, import the layout CAD model to ensure that the location of equipment objects in the factory, logistics paths, etc. are accurately consistent with the physical factory and the physical distances are equal;

[0063] Step 3. Create model objects such as equipment, buffer zone, logistics cart, data table, etc. in the simulation model in sequence;

[0064] Step 4. In the simulation model, define the logistics scheduling module and the worker scheduling module;

[0065] Step 5. In the simulation model, edit the logic script to ensure that the simulation runs according to the actual workshop process route and process time;

[0066] Step 6. In the simulation model, define the production plan verification module and build a simulation model of the tank welding workshop.

[0067] Specifically, the production plan verification module includes the following steps:

[0068] Step 1. Collect the production time and production sequence plan data of the workshop production products

[0069] Step 2. Input data into the simulation, decompose it into product parts list according to the production sequence, and set the simulation model production schedule according to the production time;

[0070] Step 3. Set the simulation model start date and work calendar;

[0071] Step 4. Run the simulation to obtain the actual completion time of each product and compare it with the planned completion time to verify the production plan.

[0072] Preferably, before step S1, the method further includes: optimizing the production sequence of each product and each part using an intelligent optimization algorithm.

[0073] Preferably, the intelligent optimization algorithm includes: GA algorithm, GA algorithm improved based on migratory bird optimization algorithm, particle swarm algorithm, taboo search algorithm, migratory bird optimization algorithm.

[0074] Preferably, when the intelligent optimization algorithm is a GA algorithm improved based on the migratory bird optimization algorithm, the optimization steps include:

[0075] A1, define the fitness function and set the number of iterations, randomly initialize the coding sequence of individuals in the population, and divide the population structure into leading birds and left and right queues.

[0076] Specifically, the optimization target is selected, and the value of the objective function is obtained after running the simulation model, such as minimizing the maximum completion time, minimizing the equipment waiting time, etc.

[0077] The coding sequence is a decimal integer sequence containing n elements, where n is the number of parts of the product; the value of the element is the number of the part type, and the position of the element in the coding sequence is the production order of the corresponding parts;

[0078] A2, after performing crossover and mutation operations on the code of each individual, a neighborhood search is performed according to the neighborhood structure, and multiple new offspring individuals are obtained by combining the sharing mechanism of other individuals;

[0079] A3, decode all individuals to obtain the production sequence of product parts, and input it into the simulation model to calculate its fitness; where the fitness function is: f = min (1 / T), T is the total completion time;

[0080] A4, sort the fitness values ​​of individuals based on the competition mechanism. After all individuals are sorted, they are divided into leading birds, left and right queues again to form a new population.

[0081] Specifically, after all individuals are sorted, the individuals with better fitness values ​​are selected and divided again into leading birds, left and right queues, to form a new population, and the population size remains unchanged.

[0082] A5, determine whether the number of iterations has been reached, if so, output the production sequence result, if not, return to step A2.

[0083] Preferably, a partial mapping crossover operator is used for the crossover operation, and an exchange mutation operator is used for the mutation operation.

[0084] Specifically, the GA algorithm improved by the migratory bird optimization algorithm optimizes the production sequence of the parts of each product produced by the tank welding workshop when they are put into production separately. The optimization goal is to minimize the maximum completion time, and the fitness function is calculated through the simulation model.

[0085] Genetic algorithm based on improved migratory bird optimization algorithm, such as Figure 2 As shown, the specific steps are as follows:

[0086] Step 1. According to the number of parts, determine the coding length, use positive integer coding as the coding method for the production order, initialize the population, and set the crossover mutation probability;

[0087] Specifically, the population initialization divides the population structure of the standard GA into three parts: the leader bird, the left and right queues. The number of the leader bird is 1, the number of the left and right queues is equal, and each part is arranged in descending order of fitness.

[0088] Step 2. Perform crossover, mutation, sharing and competition on each layer of gene sequences on the chromosome to obtain new offspring;

[0089] Among them, the crossover operation is performed inside the left and right queues or between the left and right queues, that is, the crossover operation can occur inside the left and right queues of the population, and also occurs between the left and right queues with a certain probability, which can strengthen the communication between populations, increase diversity and avoid falling into local optimality. The crossover operator adopts the partial mapping crossover method (PMX). The mutation operator adopts the exchange mutation, that is, randomly selects two genes in the sequence, exchanges their positions on the chromosomes to form new offspring individuals.

[0090] Sharing mechanism and neighborhood structure: The migratory birds in the left and right queues will receive several neighborhood solutions from the shared neighborhood solution set of the front migratory birds in the same queue. At the same time, they will search for several neighborhood solutions through the defined neighborhood structure, and select the best solution to replace them. The neighborhood structure selects insertion and greedy insertion. The insertion operation randomly selects a workpiece and randomly inserts it into the position of another chromosome; the greedy insertion operation randomly selects a workpiece. For n workpiece sequences, there will be n-1 insertion methods, and the one with the best fitness result is selected.

[0091] Fitness and competition mechanism: The fitness is calculated as the inverse of the total completion time. The shorter the completion time, the higher the fitness in the population. The total completion time is obtained based on the simulation results. The competition mechanism is to traverse the population once and sort each part of the population in descending order according to the size of the fitness; select the individual with the largest fitness to exchange positions with the leader bird.

[0092] Step 3. Decode the gene sequences of each layer on the chromosome to obtain the production order of product parts;

[0093] Step 4. Pass the parameters to the simulation software FactorySimulation to automatically adjust the production sequence parameters of the simulation model;

[0094] Step 5. Run the tank welding workshop simulation model and calculate the optimization target (minimizing the maximum completion time) as the individual fitness value;

[0095] Step 6. Use the competition mechanism to sort the fitness values ​​of individuals and select high-quality individuals to generate new parents.

[0096] Specifically, the fitness is the reciprocal of the total completion time. The shorter the completion time, the higher the fitness in the population. The total completion time is obtained based on the simulation results. The competition mechanism is that each time the population is traversed, each part of the population is sorted in descending order according to the fitness; the individual with the highest fitness is selected to exchange positions with the leading bird.

[0097] Step7. Determine whether the preset termination condition is satisfied; if satisfied, the optimization is completed, and the optimal production sequence plan for the product can be obtained after decoding the current best individual; if not satisfied, then continue to execute Step2.

[0098] S2. Arrange all products in ascending order according to the planned completion time, and let the iteration number k = 0;

[0099] S3. Let i = the total number of products;

[0100] S4. Determine whether the i-th product is overdue. If so, enter Step S4; if not, determine whether the i-th product is completed ahead of schedule. If so, enter Step S5; if not, enter Step S8;

[0101] It can be understood that the overdue time = actual completion time - planned completion time, and the early time = planned completion time - actual completion time.

[0102] S5. Let the actual production time s i = s i - t 3 , and enter Step S7; where t 3 = μ × t 1 , t 1 is the overdue time, and μ is the overdue adjustment factor;

[0103] S6. If s i + t 4 > s i+1 - T 2 , then s i = s i+1 - T 2 ; otherwise, s i = s i + t 4 , and enter Step S6; where t 4 = λ × t 2 , t 2 is the early time, and λ is the early adjustment factor;

[0104] S7. Input the updated production plan into the simulation model, run the simulation, and obtain the actual completion time of each product. If T 1 < C, end; otherwise, k = k + 1; if k % A = 0, then T 1 = T1 -B, return to step S3, otherwise, directly return to step S3; where T 1 The allowable early completion time;

[0105] S8, let i=i-1, if i>0, return to step S4, otherwise end.

[0106] Preferably, if the delay time of the i-th product is t 1 >t 1 ', then the i-th product is delayed; where, t 1 ' is the permissible delay time, t 1 '≥0.

[0107] Preferably, 0<λ<1, T 1 ≥0.

[0108] Preferably, the production plan includes the planned production start time of each product, the production sequence of each part of each product, and the production start time of each part of each product.

[0109] Specifically, the embodiment of the present invention adopts a heuristic algorithm combined with simulation iteration to find the best production plan, uses the optimized production sequence of each product component and the original production time plan of the workshop to initialize the simulation model, and finds the best production plan through a heuristic optimization method combined with simulation iteration.

[0110] This heuristic method uses a reverse sorting method. After sorting in ascending order according to the planned completion time, it gives priority to adjusting products with late delivery time, reducing the chain reaction from the front to the back and reducing the impact on the subsequent production plan. Set the minimum production interval T 2 , to avoid the production time being too intensive and the products affecting each other, the specific steps are as follows:

[0111] Step 1. Input the original production plan of the workshop and the planned completion time into the simulation model;

[0112] Step 2. Initialize the delay adjustment factor μ, the advance adjustment factor λ, and the maximum advance completion time T 1 And the minimum interval production time T 2 .

[0113] Among them, by setting the minimum production interval T 2 , to avoid the mutual influence between products due to too close production time, thus affecting the optimization effect of the production sequence of each product.

[0114] Step 3. Arrange all products in ascending order according to the planned completion time, and the number of iterations k = 0;

[0115] Step 4. Obtain the completion time of each product according to the simulation results, and let i = the total number of products;

[0116] Step 5. Determine whether the i-th product is delayed. If so, go to Step 6; determine whether the i-th product is completed ahead of schedule. If so, go to Step 7; otherwise, go to Step 9;

[0117] Specifically, the criteria for determining whether the deadline is delayed / advanced are as follows:

[0118] (1) Delay judgment criteria: Delay time t 1 >t 1 `, it is considered as delayed completion.

[0119] (2) Early completion criteria: Early completion time t 2 >T 1 , it is considered to be completed ahead of schedule.

[0120] Step 6. Calculate the delay time t 1 , calculate the adjustment time t 3 =μ×t 1 , let s i =s i -t 3 , go to Step 8;

[0121] Step 7. Calculate the lead time t 2 , calculate the adjustment time t 4 =λ×t 2 , if s i +t 4 >s i+1 -T 2 , then s i =s i+1 -T 2 Otherwise i =s i +t 4 , go to Step 8;

[0122] It is understandable that if product i is adjusted, the production interval s between product i and product i+1 i+1 -s i Less than the minimum production interval T 2 , then the production time of product i is s i+1 -T 2 ; Otherwise, the production time of product i is taken as s i .

[0123] Step 8. Update the product plan, run the plan verification module, and obtain the completion time data. k = k + 1. If k% A = 0, then T 1 =T1 -B, go to Step 4; if T 1 <10, end;

[0124] Step 9. If i>0, then i=i-1, go to Step 5; otherwise, end.

[0125] Among them, the variable maximum early completion time T is used 1 , that is, as the optimization proceeds, T 1 It is dynamically adjusted, first using a larger T 1 Perform a rough adjustment on all workpieces, and reduce T as the number of iterations increases. 1 , fine-tune the workpieces that can still be adjusted.

[0126] After many experiments, if a larger T is selected at the beginning of optimization, 1 , this heuristic method converges quickly, but the product's early completion time will just meet the requirement of less than T 1 ; If you start optimizing, take a smaller T 1 ,At this time, the requirement for the completion of each workpiece in advance is high, and it may happen that a workpiece cannot meet this constraint, making it difficult for the heuristic to converge.

[0127] Setting k%A as the judgment condition means that after A iterations, 1 The value of is reduced, that is, each product is given 1 Under the constraints of , the production time reaches the optimal optimization space.

[0128] According to the experimental results, A can be taken as 30-50, which is too small. 1 The optimization times of the value constraints are reduced, and the optimal value may not be reached. 1 If A is too large, the number of iterations will increase. B means T 1 The step size of the value change can be 5-10. If the step size is too large, the optimization result will be poor and the speed will be fast; if the step size is small, the optimization result will be good and the speed will be slow.

[0129] The method provided by the present invention is further described below with a specific example. Figure 3 As shown, it includes the following three parts:

[0130] 1. Obtain information such as workshop equipment layout, production plan, logistics methods, and process steps to build a simulation model of the tank welding workshop.

[0131] Step 1. System investigation. Conduct system investigation. The investigation content mainly includes: Equipment layout of tank welding workshop Figure 4 As shown, the process technology is as follows Figure 5 The production plan is shown in Table 1, and the product BOM is shown in Figure 6 The workshop equipment list is shown in Table 2, and the workshop resource list is shown in Table 3.

[0132] Step 2. Based on the information obtained, import the layout CAD model to ensure that the location of equipment objects in the factory, logistics paths, etc. are accurately consistent with the physical factory and the physical distances are equal;

[0133] Step 3. Create model objects such as equipment, buffer zone, logistics cart, data table, etc. in the simulation model in sequence;

[0134] Step 4. In the simulation model, define the logistics scheduling module and AGV route, set the worker scheduling module and the number of workers;

[0135] Step 5. In the simulation model, edit the logic script to ensure simulation according to the actual workshop process route and process time;

[0136] Step 6. In the simulation model, define the production plan verification module.

[0137] Specifically, the production plan verification module includes the following steps:

[0138] Step 1. Collect the production time and production sequence schedule data of the products produced in the workshop;

[0139] Step 2. Input data into the simulation, decompose it into product parts list according to the production sequence, and set the simulation model production schedule according to the production time;

[0140] Step 3. Set the simulation model start date to 2021 / 03 / 21, the work calendar to 5 days a week, and the daily working hours to 9:00-17:00;

[0141] Step 4. Run the simulation to obtain the actual completion time of each product and compare it with the planned completion time. Figure 7 As shown;

[0142] Finally, a simulation model of the tank welding workshop was constructed and the simulation was run. According to the actual production plan, 26 tanks were produced, with a start time of 2021 / 03 / 23 and a completion time of 2024 / 03 / 13, and a total time required of approximately 1112 days.

[0143] Table 1 Production plan and order

[0144]

[0145]

[0146] Table 2 Equipment list and human resource requirements

[0147] Equipment No. Equipment List Need workers 1、11 Nondestructive testing area 1, 2 3 inspectors 2 XX-4 tank docking equipment 3 general assemblers 3 XX-6 tank docking equipment 3 general assemblers 4 Material buffer and temporary storage area 5 XX-5 tank general docking equipment; 3 general assemblers 6 CNC cantilever friction stir welding 3 friction stir welding workers 7 Five-axis friction stir welding 3 friction stir welding workers 8 Vertical longitudinal seam friction stir welding 3 friction stir welding workers 9 Edge milling machine, siding pretreatment equipment; 2 CNC workers 10 CNC vertical lathe, machining common equipment 1 CNC worker 11 Test area, used for single bottom workpiece testing; 1 inspector 12 Box bottom space curve friction stir welding 3 workers for friction stir welding

[0148] Table 3 Workshop resources table

[0149] Resource Name Number of resources CNC workers 4 people Friction stir welding worker 7 people Assembly Worker 16 people Inspection Worker 4 people Number of overhead cranes 2 cars

[0150] 2. Optimization of the production sequence of product parts. The GA algorithm improved based on the migratory bird optimization algorithm optimizes the production sequence of all product parts in the tank welding workshop when they are put into production separately, and the fitness function is calculated through the simulation model.

[0151] Step 1. According to the number of parts, determine the length of the code, and use the positive integer coding method as the coding for the production sequence to initialize the population; this workshop produces a total of 3 products, namely XX-4, XX-5 and XX-6, refer to Figure 4 According to the product BOM diagram, XX-4 tank requires 36 parts, XX-5 tank requires 42 parts, and XX-6 tank requires 52 parts. Therefore, when optimizing the production sequence of XX-4 tank, the code length is set to 36, when optimizing the production sequence of XX-5 tank, the code length is set to 42, and when optimizing the production sequence of XX-6 tank, the code length is set to 52.

[0152] The population size is set to 51. The population initialization divides the population structure of the standard GA into three parts: the leader bird, the left and right queues. The number of the leader bird is 1, and the number of the left and right queues is equal to 25. Each part is arranged in descending order of fitness. The crossover mutation probabilities are set to 0.8 and 0.1 respectively.

[0153] Step 2. Perform crossover, mutation, sharing and competition on each layer of gene sequences on the chromosome to obtain new offspring;

[0154] Among them, the crossover operation can occur within the left and right queues of the population, and also occurs between the left and right queues with a certain probability, which can strengthen the communication between populations, increase diversity and avoid falling into local optimality. The crossover operator adopts the partial mapping crossover method (PMX). The mutation operator adopts the exchange mutation, that is, randomly selects two genes in the sequence, exchanges their positions on the chromosomes, and forms new offspring individuals.

[0155] Sharing mechanism and domain structure: The migratory birds in the left and right queues will receive several neighborhood solutions from the shared neighborhood solution set of the front migratory birds in the same queue. At the same time, they will search for several neighborhood solutions through the defined neighborhood structure, and select the best solution to replace them. The domain structure selects insertion and greedy insertion methods. The insertion operation randomly selects a workpiece and randomly inserts it to another position; the greedy insertion operation randomly selects a workpiece. For a sequence of n workpieces, there will be n-1 insertion methods, and the one with the best fitness result will be selected.

[0156] Step 3. Decode the gene sequences of each layer on the chromosome to obtain the production order of product parts;

[0157] Step 4. Pass the parameters to the simulation software to automatically adjust the production sequence parameters of the simulation model;

[0158] Step 5. Run the tank welding workshop simulation model to calculate the parameter value of the optimization target as the individual fitness value;

[0159] Step 6. Use the competition mechanism to sort the fitness values ​​of individuals and select high-quality individuals to generate new parents.

[0160] Specifically, the fitness is calculated as the inverse of the total completion time. The shorter the completion time, the higher the fitness in the population. The total completion time is obtained based on the simulation results. The competition mechanism is to traverse the population once, sort each part of the population in descending order according to the size of the fitness; select the individual with the largest fitness to exchange positions with the leader bird.

[0161] Step 7. Determine whether the preset termination condition is met; if so, the optimization is completed, and the optimal production sequence plan for the product can be obtained after decoding the current best individual; if not, continue to execute Step 2.

[0162] Optimization results: The production time of XX-4, 5, and 6 before optimization was 64 days and 12 hours, 63 days and 15 hours, and 76 days and 15 hours respectively; after optimization, the production time of XX-4, 5, and 6 was 58 days and 14 hours, 59 days and 11 hours, and 66 days and 10 hours respectively.

[0163] 3. Use the optimized production sequence of the above product parts and the original production time plan of the workshop to initialize the simulation model, and use the heuristic optimization method combined with simulation iteration to find the best production plan. The specific steps are as follows.

[0164] Step 1. Input the actual production time plan and planned completion time of the workshop into the simulation, such as Figure 7 As shown;

[0165] Step 2. Initialize the delay adjustment factor μ to 0.5, the advance adjustment factor λ to 1, and the maximum early completion time T 1 90 days and the minimum interval time between commissioning is T 2 8 days;

[0166] Step 3. Arrange all products in ascending order according to the planned completion time, and the number of iterations k = 0;

[0167] Step 4. Obtain the completion time of each product according to the simulation results, and let i = the total number of product orders;

[0168] Step 5. Determine whether the i-th product is delayed. If so, go to Step 6; determine whether the i-th product is completed ahead of schedule. If so, go to Step 7; otherwise, go to Step 9;

[0169] Specifically, the criteria for determining whether the deadline is delayed / advanced are as follows:

[0170] (1) Delay judgment criteria: Delay time t 1 >t 1 `, it is considered as delayed completion.

[0171] (2) Early completion criteria: Early completion time t 2 >T 1 , it is considered to be completed ahead of schedule.

[0172] Step 6. Calculate the delay time t 1 , calculate the adjustment time t 3 =μ×t 1 , let s i =s i -t 3 , go to Step 8;

[0173] Step 7. Calculate the lead time t 2 , calculate the adjustment time t 4 =λ×t 2 , if s i +t 4 >s i+1 -T 2 , then s i =s i+1 -T 2 Otherwise i =s i +t 4 , go to Step 8;

[0174] Step 8. Update the product plan, run the plan verification module, and obtain the completion time data. k = k + 1. If k% 30 = 0, then T 1 =T 1 -5, go to Step 4; if T 1 <10, end (i.e., A=30, B=5);

[0175] Step 9: If i>0, then i=i-1, go to Step 5; otherwise, end;

[0176] The average early completion time was 90.7 days, and the total completion time was 895 days, which was a significant improvement over the initial plan. The results before optimization were as follows: Figure 8 The optimized results are shown in Fig. 9As shown in the figure, the product launch plan meets the optimization goal of obtaining a later launch time without allowing delays, which verifies the effectiveness of this method.

[0177] The simulation-based tank welding workshop production planning optimization system provided by the present invention is described below.

[0178] An embodiment of the present invention provides a simulation-based tank welding workshop production plan optimization system, including: a computer-readable storage medium and a processor;

[0179] The computer-readable storage medium is used to store executable instructions;

[0180] The processor is used to read the executable instructions stored in the computer-readable storage medium and execute the method described in any of the above embodiments.

[0181] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing the production plan of a tank welding workshop based on simulation, characterized in that: include: S1, establish a tank welding workshop planning optimization simulation model, input the production plan and planned completion time into the simulation model, and obtain the actual completion time of each product; S2, sort all products in ascending order according to the planned completion time, and set the number of iterations k = 0; S3, let i = total number of products; S4, determine whether the i-th product is overdue, if yes, proceed to step S5; if no, determine whether the i-th product is completed ahead of schedule, if yes, proceed to step S6, if no, proceed to step S8; S5, the actual production time of the product is s i =s i -t3, go to step S7; wherein t3 = μ×t1, t1 is the delay time, μ is the delay adjustment factor; S6, if s i +t4>s i+1 -T2, then s i =s i+1 -T2; otherwise, s i =s i +t4, enter step S7; wherein t4=λ×t2, t2 is the advance time, λ is the advance adjustment factor, and T2 is the minimum production interval; S7, input the updated production plan into the simulation model, obtain the actual completion time of each product, if T1<C, end; otherwise, k=k+1; if k%A=0, then T1=T1-B, return to step S3, otherwise, directly return to step S3; wherein T1 is the allowable early completion time; S8, let i=i-1, if i>0, return to step S4, otherwise end; If the delay time of the i-th product t1>t1`, the i-th product is delayed; wherein t1` is the allowable delay time, t1`≥0; 0<λ<1, T1≥0; The method for establishing the tank welding workshop planning optimization simulation model is as follows: B1, import the equipment layout, process technology, production plan, workshop equipment and resource list of the tank welding workshop into the layout CAD model; B2, create equipment, buffer, logistics cart, data table model objects in sequence; define logistics scheduling module and worker scheduling module; B3 edits the logic script in the simulation model to ensure that the simulation runs according to the actual workshop process route and process time; B4, define the production plan verification module: collect the production time and production sequence plan data of the products produced in the workshop; input the data into the simulation, decompose it into product parts list according to the production sequence, and set the simulation model production plan according to the production time; set the simulation model start date and work calendar; run the simulation, obtain the actual completion time of each product, and compare it with the planned completion time.

2. The method according to claim 1, characterized in that The production plan includes the planned production time of each product, the production sequence of each part of each product, and the production time of each part of each product.

3. The method according to claim 1, characterized in that Before step S1, the method further includes: optimizing the production sequence of each part of each product by using an intelligent optimization algorithm.

4. The method according to claim 3, characterized in that The intelligent optimization algorithms include: GA algorithm, particle swarm algorithm, taboo search algorithm, and migratory bird optimization algorithm.

5. The method according to claim 4, characterized in that The GA algorithm includes a GA algorithm improved based on a migratory bird optimization algorithm.

6. The method according to claim 5, characterized in that When the intelligent optimization algorithm is a GA algorithm improved based on the migratory bird optimization algorithm, the optimization steps include: A1, define the fitness function and set the number of iterations, randomly initialize the coding sequence of individuals in the population, and divide the population structure into leading birds and left and right queues; The coding sequence is a decimal integer sequence containing n elements, where n is the number of parts of the product; the value of the element is the number of the part type, and the position of the element in the coding sequence is the production order of the corresponding parts; A2, after performing crossover and mutation operations on the codes of each individual, a neighborhood search is performed according to the neighborhood structure, and multiple new offspring individuals are obtained by combining the sharing mechanism of other individuals; A3, decode each individual to obtain the production sequence of product parts, and input it into the simulation model to calculate its fitness; where the fitness function is: f = min (1 / T), T is the total completion time; A4, sort the fitness values ​​of individuals based on the competition mechanism. After all individuals are sorted, they are divided into leading birds, left and right queues, and a new population is formed; A5, determine whether the number of iterations has been reached, if so, output the optimal production sequence result, if not, return to step A2.

7. The method according to claim 6, characterized in that The partial mapping crossover operator is used for crossover operation, and the swap mutation operator is used for mutation operation; The neighborhood structure adopts selective insertion and greedy insertion.

8. A simulation-based tank welding workshop production planning optimization system, characterized in that: include: A computer readable storage medium and a processor; The computer-readable storage medium is used to store executable instructions; The processor is used to read the executable instructions stored in the computer-readable storage medium and execute the method according to any one of claims 1 to 7.

Images