A process route coordinated general assembly production process method

Abstract

The application discloses a kind of process route coordination's general assembly production process method, comprising the following steps: process route coordination's general assembly production process problem is described, and decision variable is set;Process route mathematical model analysis is carried out to the constraint condition of general assembly production process problem;Analysis optimization target, constructs target optimization function;Design an improved self-adapting genetic algorithm to solve target optimization function.The present application solves the problems of existing technologies such as inaccurate process coordination modeling, insufficient multiple constraint coordination capability, poor dynamic adaptability of resources, mismatch between optimization algorithm and actual problem structure, and low degree of visualization of decision support, thereby significantly improving process scheduling accuracy and feasibility, improving production efficiency, optimizing resource utilization, significantly improving production balance, enhancing decision support capability, and significantly improving algorithm solving efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent production process technology in the high-end equipment manufacturing industry, and in particular to a process route coordination method for final assembly production. Background Technology

[0002] In the high-end equipment manufacturing industry, final assembly production faces unprecedented complexity and challenges. Traditional production scheduling and process routes are no longer sufficient to meet modern production needs, and the main technical bottlenecks are as follows:

[0003] Inaccurate process collaboration modeling: Traditional production methods use simplified linear process network models, which cannot accurately depict the complex collaborative relationships unique to final assembly production, such as multi-stage assembly, parallel integration of final assembly and inspection, and strict process dependencies. Existing models cannot accurately express these spatiotemporally coupled process collaboration characteristics, leading to a serious disconnect between process routing schemes and actual production; Insufficient coordination capability for multiple constraints: Production workshops face multiple hard constraints such as low buffer capacity, component temperature settling, and workshop staffing. Existing process routing methods typically handle various constraints independently, lacking a systematic mechanism for handling coupled constraints, making it difficult to generate feasible solutions that simultaneously meet all constraints; Poor dynamic resource adaptability: Traditional production uses static resource allocation strategies, which cannot adapt to the dynamic changes in resource requirements between processes. Final assembly production involves specialized assembly teams and inspection teams, and each process has different requirements for resource types, quantities, and time. Existing methods struggle to achieve dynamic optimization of resource allocation, leading to low resource utilization and significant non-productive time losses due to frequent job switching. Furthermore, optimization algorithms are mismatched with the actual problem structure: existing intelligent optimization algorithms suffer from structural defects when applied to final assembly production. Standard genetic algorithms' single-layer encoding cannot simultaneously represent process sequence and resource allocation; fixed-parameter strategies are ill-suited to the needs of different stages of the search process; and there is a lack of dedicated optimization operators specific to the characteristics of final assembly production. These defects result in slow algorithm convergence and poor solution quality, failing to meet the dual requirements of process scheduling efficiency and solution quality in actual production. Finally, decision support visualization is low: current production systems primarily output data reports, lacking intuitive spatiotemporal relationship visualization. Production managers struggle to quickly grasp the overall process scheduling logic, identify resource conflicts, and assess solution feasibility from massive amounts of data, severely impacting the actual executability and adjustment efficiency of process scheduling solutions. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a final assembly production process method with coordinated process routes;

[0005] The technical solution of this invention to solve the above-mentioned technical problems is: a final assembly production process method with coordinated process routes; comprising the following steps:

[0006] Step S1: Describe the assembly production process problem of process route coordination and set decision variables;

[0007] Step S2: Perform mathematical model analysis of the process route for the constraints of the final assembly production process problem; the constraints of the final assembly production process problem include buffer capacity constraints, human resource coordination constraints, temperature settling and control component capacity constraints, process route sequence constraints, batch logic constraints, and other basic constraints.

[0008] Step S3: Analyze the objective to be optimized and construct the corresponding objective optimization function;

[0009] Step S4: Design an improved adaptive genetic algorithm to solve the objective optimization function; the specific process of step S4 is as follows:

[0010] S41, establish a two-layer coding mechanism;

[0011] S42, Fitness Assessment and Constraint Handling;

[0012] S43, an adaptive strategy based on the population convergence state dynamically adjusts the crossover probability and mutation probability;

[0013] S44 is a local optimization operator that reduces job switching.

[0014] In the above-mentioned assembly production process method with coordinated process routes, step S1 describes the assembly production process problem as follows:

[0015] Suppose we are manufacturing core components for high-precision products, denoted as h. The total task comprises N sets of products. We divide the total task into K production batches, satisfying the following conditions: Where k = 1, 2, ..., K Let X be the set of indexes of all products included in the k-th production batch; let t be the total task set; let t be the time in days; the processing of each product is decomposed into 10 processes, which include 5 assembly processes and 5 inspection processes. The 5 assembly processes are denoted as A1, A2, A3, A4, and A5, and the set of assembly processes is... The five inspection procedures are denoted as I1, I2, I3, I4, and I5, and the set of inspection procedures is as follows: The operator team is composed of The inspectors assembled Both the operator team and the inspector are considered human resources, denoted as r, where r belongs to... or Buffer zone: Maximum capacity of 160 sets, considered as space resource.

[0016] In the above-mentioned assembly production process method with coordinated process routes, the decision variables set in step S1 are as follows:

[0017] This indicates the processing status of the j-th process of the i-th product on day t, where i represents the i-th product and j represents the j-th process. If the j-th process of the i-th product begins processing on day t... =1, otherwise =0, ;

[0018] : Indicates the start time of the j-th process of the i-th product, in days;

[0019] : Represents the completion time of the j-th process of the i-th product, in days;

[0020] : This indicates the resource occupancy status of the j-th process of the i-th product set using human resources r. If the j-th process of the i-th product set uses either the operator team or the inspector, then... =1, otherwise =0, ;

[0021] : Represents the spatial state of the i-th product on day t. If the i-th product is in the buffer zone on day t, then... =1, otherwise =0, .

[0022] In the above-mentioned assembly production process method with coordinated process routes, the constraints of the assembly production process problem in step S2 are as follows:

[0023] Buffer capacity constraint: The number of buffer sets occupied by work-in-process shall not exceed 160 sets;

[0024] ;

[0025] in, This indicates that the i-th product is in the buffer zone on day t, meaning that there exists a j-th process that makes... , for The immediate following process; For the i-th set of products The start time of each process;

[0026] Human resource coordination constraints include operator constraints, inspector constraints, and workshop staffing constraints.

[0027] Operator constraints:

[0028] ;

[0029] in, It is a function of time t, when When, the function value is 1, when When the time is right, the function value is 0;

[0030] The operator constraint means that each operator group can undertake a maximum of one assembly task at a time.

[0031] Inspector constraints:

[0032] ;

[0033] The inspector constraint means that each inspector can undertake a maximum of one inspection task at a time;

[0034] Workshop staffing constraints:

[0035] ;

[0036] Workshop staffing constraints are used to ensure that the total number of people in the workshop does not exceed 9.

[0037] Temperature settling and control component capacity constraints include the first assembly process start time constraint and the daily capacity constraint of core control components.

[0038] For core components requiring constant temperature, the start time of the first assembly process must satisfy the following formula:

[0039] ;

[0040] in, The warehousing time of the core component h of the i-th product is in days. This is the first assembly process for core components that require constant temperature. The first assembly process for the core components of the i-th product that require constant temperature. The start time, Three different sub-components representing the core components of Class A. Three different sub-components representing the core components of Class B;

[0041] Controlling the daily production capacity constraints of core components:

[0042] ;

[0043] in, This refers to the set of assembly processes for core components; the daily production capacity constraint for core components indicates that the daily completion quantity of related processes for core components is less than or equal to 40.

[0044] Process route sequence constraints: A subsequent process can only begin after all preceding processes have been completed. The specific process route description includes process route construction, parallel paths, merging points, branch synchronization, and final inspection.

[0045] Process route construction: Establish serial and parallel relationships between processes, including two parallel paths A1→I1→A3→I3→A5→I5 and A2→I2→A3→I3→A5→I5, as well as an independent path A4→I4→A5→I5, forming a process network structure;

[0046] Parallel paths: A1→I1 and A2→I2 can be parallelized;

[0047] Meeting point: A3 can only begin after I1 and I2 are completed;

[0048] Branch synchronization: A3→I3 and A4→I4 are performed in parallel. A5 can only begin after both are completed.

[0049] Final verification: After A5 is completed, proceed to I5;

[0050] Among them, the process cannot be preempted, each stage includes mandatory inspection, and the inspection is the immediate preceding process of the next assembly;

[0051] ;

[0052] in, Let A be the completion time of assembly process A1 for the i-th product. Let I be the start time of the inspection process I1 for the i-th set of products. Let I be the completion time of inspection process I1 for the i-th set of products. Let be the completion time of assembly process A2 for the i-th product. Let I be the start time of inspection process I2 for the i-th set of products. Let I be the completion time of inspection process I2 for the i-th set of products. Let A be the start time of assembly process A3 for the i-th product. Let A be the completion time of assembly process A3 for the i-th product. Let I be the start time of inspection process I3 for the i-th set of products. Let I be the completion time of inspection process I3 for the i-th set of products. Let A5 be the start time of the assembly process A5 for the i-th product. Let A be the completion time of assembly process A4 for the i-th product. Let I be the start time of inspection process I4 for the i-th set of products. Let I be the completion time of inspection process I4 for the i-th set of products. Let A5 be the completion time of the assembly process A5 for the i-th product. Let I be the start time of inspection process I5 for the i-th set of products;

[0053] Batch logic constraint: Products are produced sequentially, and the production line must be cleared between batches;

[0054] ;

[0055] in, To minimize the function, To maximize the function, Let this be the start time of the first process for the i-th product. Let M be the completion time of the final process for the i-th product, and M be the total number of processes in the process route. This is the set of indexes for all products included in the (k+1)th production batch.

[0056] The batch logic constraint means that the earliest start time of the first process of any product in the (k+1)th production batch must be later than or equal to the latest completion time of the last process of all products in the kth production batch.

[0057] Other fundamental constraints include process time and exclusive human resources;

[0058] Process time: , For process Standard working hours;

[0059] Exclusive access to human resources: Each process requires only one type of human resource.

[0060] The specific process of step S3 in the above-mentioned assembly production process method with coordinated process routes is as follows:

[0061] Step S31, Optimization Target Analysis: Based on Step S1 and Step S2, the core optimization target is defined as minimizing the total production cycle;

[0062] Step S32, Objective Function Construction: Construct the objective optimization function that minimizes the total production cycle, with the mathematical expression as follows:

[0063] ;

[0064] in, Indicates the total production cycle. Let I be the completion time of inspection process I5 for the i-th set of products. The objective function is to minimize the total production cycle.

[0065] The specific process of step S41 in the above-mentioned assembly production process method with coordinated process routes is as follows:

[0066] First, the upper-level process sequence is encoded: process-based encoding is used, and the chromosome is represented as a first vector. , , Let L be the L-th gene in the first vector G, where L is the first vector. The length of L is equal to the total number of processes. Let y be a binary tuple (y, d), where y is the product number and d is the process type number, and d = 1, 2, ..., 10 correspond to 5 assembly processes and 5 inspection processes respectively;

[0067] Then, the lower-level resource allocation coding is performed: a human resource-based coding method is adopted, and the chromosome is represented as a second vector. , It is the same length as G; For the Lth gene in the second vector F, This indicates the resource number assigned to the corresponding process. For assembly processes, then ∈ , indicating one of the operator teams, if To inspect the process, then ∈ , indicating one of the inspectors.

[0068] The specific process of step S42 in the above-mentioned assembly production process method with coordinated process routes is as follows:

[0069] The fitness is calculated based on the objective function, using the following formula:

[0070] ;

[0071] Where λ is the penalty coefficient. For fitness; P is the penalty for constraint violation;

[0072] The formula for calculating the penalty item P for constraint violation is:

[0073] ;

[0074] in, for The weighting coefficients, for The weighting coefficients, Let Q be the weighting factor for Q, where Q represents the violation of other constraints. Let t be the number of product sets that are in the buffer zone and have not been transferred on day t. Let t be the number of people present in the workshop at the same time on day t.

[0075] The constraint handling strategy is as follows:

[0076] Hard constraint handling: Hard constraints such as process sequence and temperature settling are forcibly satisfied during the decoding process;

[0077] Soft constraint handling: Buffer capacity, workshop staffing, etc. are handled through penalty functions to guide the search towards convergence to the feasible region;

[0078] Repair strategy: For slightly infeasible solutions, heuristic methods such as time shifting and resource reallocation are used for repair.

[0079] The specific process of step S43 in the above-mentioned assembly production process method with coordinated process routes is as follows:

[0080] To address the shortcomings of fixed parameters in traditional genetic algorithms, an adaptive parameter adjustment strategy based on the population convergence state is proposed:

[0081] The first step, convergence state measurement: calculate the standard deviation of the current population fitness. Standard deviation of fitness from the initial population By comparison, the convergence index δ is obtained, δ= / ;

[0082] The second step is to calculate the correlation probability of the upper-level process sequence encoding: the first crossover probability. and the first mutation probability Adjust adaptively according to the following formula;

[0083] ;

[0084] ;

[0085] in, This is the preset lower bound for the first crossover probability. The upper limit of the first crossover probability is preset. This is the preset lower limit of the first mutation probability. This is the preset upper limit of the first mutation probability. This is the adjustment coefficient for the first crossover probability. This is the adjustment coefficient for the first mutation probability;

[0086] Step 3: Population diversity assessment: Calculate the standard deviation of the current population fitness in each generation of evolution. , Quantitatively characterize the degree of population dispersion;

[0087] Step 4: Calculate the relevant probabilities of lower-level resource allocation codes and dynamically adjust the probabilities: based on The size of the second crossover probability is dynamically adjusted. Second mutation probability ;

[0088] Dynamic adjustment is performed using a nonlinear dynamic adjustment formula, which is as follows:

[0089] ;

[0090] ;

[0091] in, This is the base value for the second crossover probability. This is the base value for the second mutation probability. is the preset adaptive adjustment coefficient, and e is the base of the natural logarithm.

[0092] The specific process of step S44 in the above-mentioned assembly production process method with coordinated process routes is as follows:

[0093] First, considering the need for collaborative work among operators, a job switching optimization operator is designed:

[0094] when At the same time, for each operator group, calculate the number of job changes in its process sequence. ;choose The largest group is selected as the optimization target; all processes responsible for the optimization target are extracted and recorded as a process sequence list. , , For list The m-th process in the list, where m is the number of steps in the list. The length of , where Includes process type, start time, and product information; under the premise of meeting process sequence constraints, processes of the same type are arranged according to the principle of adjacent processes. The sorting algorithm, with the core objective of minimizing the number of type switching between adjacent processes, constructs an objective function to group processes of the same type together, thereby reducing frequent switching between different types of processes on the production line. The objective function is:

[0095] ;

[0096] in, To minimize the function, Here, m represents the convergence index, and m is a list. length, Indicate extraction The process type of the j-th process. Indicate extraction The process type of the (j+1)th process; when hour Take 1, otherwise Set to 0;

[0097] Then, maintaining the dependencies between processes, recalculate the start and finish times of each process after the reorganization, verify whether the new solution meets all constraints, and replace the original solution if it does.

[0098] The above method also includes: implementing the algorithm using MATLAB software and solving it by programming, outputting a resource load map, the specific process of which is as follows;

[0099] Algorithm implementation and solution steps:

[0100] The improved adaptive genetic algorithm described above is implemented using the MATLAB programming platform. The objective optimization function constructed in step S3 and the constraints established in step S2 are converted into a computable mathematical model. Algorithm parameters are set, including population size, maximum number of iterations, mutation probability range, and local optimization probability. Iterative calculations of the genetic algorithm are executed, evolving generation by generation until the termination condition is met through operations such as population initialization, fitness calculation, crossover mutation, and local optimization. The optimal chromosome code is output, and decoded to generate a complete production process route plan, including product information such as the start time, end time, and resource allocation for each process.

[0101] Visual output generation steps:

[0102] Generate a resource load diagram, with the horizontal axis representing time and the vertical axis representing resource load status. Plot the workload curves of the assembly team and the inspectors respectively, showing the busy and idle status of each resource at different time periods.

[0103] Generate a daily production completion chart. Daily production statistics are based on the completion time of the final inspection process. The chart shows the number of products that complete the final inspection each day. The horizontal axis represents the number of production days, and the vertical axis represents the number of sets produced per day.

[0104] Generate a buffer usage graph to show the buffer capacity usage, reflecting the buffer occupancy and product utilization.

[0105] The beneficial effects of this invention are as follows:

[0106] 1. Significantly improved accuracy and feasibility of process scheduling: Through an improved adaptive genetic algorithm combined with collaborative process modeling and coupling constraint processing, the generated process scheduling schemes closely match actual production. The feasibility of the schemes is greatly improved, reducing on-site adjustments and enhancing the executability of the process scheduling schemes.

[0107] 2. Significantly Improved Production Efficiency: By applying the method of this invention, the total production cycle is significantly shortened, and production efficiency is effectively improved. Taking the final assembly of high-precision products as an example, the average daily output capacity is enhanced, the capacity utilization rate is increased, and the contract delivery capability is significantly improved.

[0108] 3. Significantly improved resource utilization: Through dual-layer intelligent coding and dynamic resource adaptation, the overall resource utilization rate has been significantly optimized. The assembly team's work schedule is more rational, the inspector's workload is more balanced, the number of job switching frequency is reduced, and non-productive time loss is decreased.

[0109] 4. Significantly improved production balance: Buffer zone occupancy is more balanced, and production rhythm is more stable. The utilization of special resources such as control components is more rational, ensuring the stability and continuity of the production process.

[0110] 5. Significantly improved algorithm efficiency: Compared to traditional algorithms, the improved adaptive genetic algorithm converges faster and produces higher-quality optimal solutions. The time required to generate process scheduling plans is greatly reduced, enabling real-time adjustments to meet the needs of the production site.

[0111] 6. Enhanced Decision Support Capabilities: The multi-dimensional visualization output system transforms complex process scheduling logic into an intuitive visual presentation, enabling production managers to quickly grasp the overall process scheduling plan, identify potential conflicts, and evaluate the merits of different plans. Decision response time is shortened, and adjustment efficiency is improved.

[0112] 7. Significant economic benefits: Applying the method of this invention can effectively increase production capacity, save resource costs, shorten delivery cycle, and reduce work-in-process inventory, bringing significant economic benefits to enterprises and enhancing their market competitiveness;

[0113] In summary, this invention fills the technological gap in intelligent scheduling for final assembly production in the field of high-end equipment manufacturing. Through innovative process collaborative modeling, coupled constraint processing, dual-layer intelligent coding, adaptive optimization algorithms, and multi-dimensional visualization technology, it achieves efficient scheduling of final assembly production lines under multiple constraints, providing the industry with a practical solution and possessing significant engineering application value and promising prospects for promotion. Attached Figure Description

[0114] Figure 1 This is a flowchart of the core steps of the present invention;

[0115] Figure 2 This is a flowchart of the improved adaptive genetic algorithm of this invention;

[0116] Figure 3 This invention is a resource load diagram representing the working status of an assembly team;

[0117] Figure 4 This invention is a resource load diagram representing the working status of the inspection team;

[0118] Figure 5 This is a diagram showing the daily output of this invention.

[0119] Figure 6 This is a diagram showing the usage of the buffer in this invention. Detailed Implementation

[0120] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0121] like Figure 1As shown, this invention provides a final assembly production process method with coordinated process routes, specifically including the following core steps:

[0122] Step S1: Describe the assembly production process problem of process route coordination and set decision variables;

[0123] The final assembly production process problem is described as follows: Assume the production of core components for high-precision products is carried out, with the core component denoted as h. The total task comprises N sets of products. Divide the total task into K production batches, and satisfy the following conditions: Where k = 1, 2, ..., K Let X be the set of indexes of all products included in the k-th production batch; let t be the total task set; let t be the time in days; the processing of each product is decomposed into 10 processes, which include 5 assembly processes and 5 inspection processes. The 5 assembly processes are denoted as A1, A2, A3, A4, and A5, and the set of assembly processes is... The five inspection procedures are denoted as I1, I2, I3, I4, and I5, and the set of inspection procedures is as follows: The operator team is composed of The inspectors assembled Both the operator team and the inspector are considered human resources, denoted as r, where r belongs to... or Buffer zone: Maximum capacity of 160 sets, considered as space resource.

[0124] The decision variables are set as follows:

[0125] This indicates the processing status of the j-th process of the i-th product on day t, where i represents the i-th product and j represents the j-th process. If the j-th process of the i-th product begins processing on day t... =1, otherwise =0, ;

[0126] : Indicates the start time of the j-th process of the i-th product, in days;

[0127] : Represents the completion time of the j-th process of the i-th product, in days;

[0128] : This indicates the resource occupancy status of the j-th process of the i-th product set using human resources r. If the j-th process of the i-th product set uses either the operator team or the inspector, then... =1, otherwise =0, ;

[0129] : Represents the spatial state of the i-th product on day t. If the i-th product is in the buffer zone on day t, then... =1, otherwise =0, .

[0130] Step S2: Perform mathematical model analysis of the process route for the constraints of the final assembly production process problem; the constraints of the final assembly production process problem include buffer capacity constraints, human resource coordination constraints, temperature settling and control component capacity constraints, process route sequence constraints, batch logic constraints, and other basic constraints.

[0131] Buffer capacity constraint: The number of buffer sets occupied by work-in-process shall not exceed 160 sets;

[0132] ;

[0133] in, This indicates that the i-th product is in the buffer zone on day t, meaning that there exists a j-th process that makes... , for The immediate following process; For the i-th set of products The start time of each process;

[0134] Human resource coordination constraints include operator constraints, inspector constraints, and workshop staffing constraints.

[0135] Operator constraints:

[0136] ;

[0137] in, It is a function of time t, when When, the function value is 1, when When the time is right, the function value is 0;

[0138] The operator constraint means that each operator group can undertake a maximum of one assembly task at a time.

[0139] Inspector constraints:

[0140] ;

[0141] The inspector constraint means that each inspector can undertake a maximum of one inspection task at a time;

[0142] Workshop staffing constraints:

[0143] ;

[0144] Workshop staffing constraints are used to ensure that the total number of people in the workshop does not exceed 9.

[0145] Temperature settling and control component capacity constraints include the first assembly process start time constraint and the daily capacity constraint of core control components.

[0146] For core components requiring constant temperature, the start time of the first assembly process must satisfy the following formula:

[0147] ;

[0148] in, Let h be the warehousing time of the core component h of the i-th product, in days. In the formula, "+1" indicates that the core component requiring constant temperature needs to be left to stand for 24 hours. This is the first assembly process for core components that require constant temperature. The first assembly process for the core components of the i-th product that require constant temperature. The start time, Three different sub-components representing the core components of Class A. Three different sub-components representing the core components of Class B;

[0149] Controlling the daily production capacity constraints of core components:

[0150] ;

[0151] in, This refers to the set of assembly processes for core components; the daily production capacity constraint for core components indicates that the daily completion quantity of related processes for core components is less than or equal to 40.

[0152] Process route sequence constraints: A subsequent process can only begin after all preceding processes have been completed. The specific process route description includes process route construction, parallel paths, merging points, branch synchronization, and final inspection.

[0153] Process route construction: Establish serial and parallel relationships between processes, including two parallel paths A1→I1→A3→I3→A5→I5 and A2→I2→A3→I3→A5→I5, as well as an independent path A4→I4→A5→I5, forming a process network structure;

[0154] Parallel paths: A1→I1 and A2→I2 can be parallelized;

[0155] Meeting point: A3 can only begin after I1 and I2 are completed;

[0156] Branch synchronization: A3→I3 and A4→I4 are performed in parallel. A5 can only begin after both are completed.

[0157] Final verification: After A5 is completed, proceed to I5;

[0158] Among them, the process cannot be preempted, each stage includes mandatory inspection, and the inspection is the immediate preceding process of the next assembly;

[0159] ;

[0160] in, Let A be the completion time of assembly process A1 for the i-th product. Let I be the start time of the inspection process I1 for the i-th set of products. Let I be the completion time of inspection process I1 for the i-th set of products. Let be the completion time of assembly process A2 for the i-th product. Let I be the start time of inspection process I2 for the i-th set of products. Let I be the completion time of inspection process I2 for the i-th set of products. Let A be the start time of assembly process A3 for the i-th product. Let A be the completion time of assembly process A3 for the i-th product. Let I be the start time of inspection process I3 for the i-th set of products. Let I be the completion time of inspection process I3 for the i-th set of products. Let A5 be the start time of the assembly process A5 for the i-th product. Let A be the completion time of assembly process A4 for the i-th product. Let I be the start time of inspection process I4 for the i-th set of products. Let I be the completion time of inspection process I4 for the i-th set of products. Let A5 be the completion time of the assembly process A5 for the i-th product. Let I be the start time of inspection process I5 for the i-th set of products;

[0161] Batch logic constraint: Products are produced sequentially, and the production line must be cleared between batches;

[0162] ;

[0163] in, To minimize the function, To maximize the function, Let this be the start time of the first process for the i-th product. Let M be the completion time of the final process for the i-th product, and M be the total number of processes in the process route. This is the set of indexes for all products included in the (k+1)th production batch.

[0164] The batch logic constraint means that the earliest start time of the first process of any product in the (k+1)th production batch must be later than or equal to the latest completion time of the last process of all products in the kth production batch.

[0165] Other fundamental constraints include process time and exclusive human resources;

[0166] Process time: , For process Standard working hours;

[0167] Exclusive access to human resources: Each process requires only one type of human resource.

[0168] Step S3: Analyze the objective to be optimized and construct the corresponding objective optimization function;

[0169] The specific process of step S3 is as follows:

[0170] Step S31, Optimization Target Analysis: Based on Step S1 and Step S2, the core optimization target is defined as minimizing the total production cycle;

[0171] Step S32, Objective Function Construction: Construct the objective optimization function that minimizes the total production cycle, with the mathematical expression as follows:

[0172] ;

[0173] in, Indicates the total production cycle. Let I be the completion time of inspection process I5 for the i-th set of products. The objective function is to minimize the total production cycle.

[0174] Step S4: Design an improved adaptive genetic algorithm to solve the objective optimization function;

[0175] like Figure 2 As shown, the specific process of step S4 is as follows: S41, establish a two-layer coding mechanism;

[0176] First, the upper-level process sequence is encoded: process-based encoding is used, and the chromosome is represented as a first vector. , , Let L be the L-th gene in the first vector G, where L is the first vector. The length of L is equal to the total number of processes. Let y be a binary tuple (y, d), where y is the product number and d is the process type number, and d = 1, 2, ..., 10 correspond to 5 assembly processes and 5 inspection processes respectively;

[0177] Then, the lower-level resource allocation coding is performed: a human resource-based coding method is adopted, and the chromosome is represented as a second vector. , It is the same length as G; For the Lth gene in the second vector F, This indicates the resource number assigned to the corresponding process. For assembly processes, then ∈ , indicating one of the operator teams, if To inspect the process, then ∈ , indicating one of the inspectors.

[0178] S42, Fitness Assessment and Constraint Handling;

[0179] The fitness is calculated based on the objective function, using the following formula:

[0180] ;

[0181] Where λ is the penalty coefficient. For fitness; P is the penalty for constraint violation.

[0182] The formula for calculating the penalty item P for constraint violation is:

[0183] ;

[0184] in, for The weighting coefficients, for The weighting coefficients, Let Q be the weighting factor for Q, where Q represents the violation of other constraints. Let t be the number of product sets that are in the buffer zone and have not been transferred on day t. Let t be the number of people present in the workshop at the same time on day t.

[0185] The constraint handling strategy is as follows:

[0186] Hard constraint handling: Hard constraints such as process sequence and temperature settling are forcibly satisfied during the decoding process;

[0187] Soft constraint handling: Buffer capacity, workshop staffing, etc. are handled through penalty functions to guide the search towards convergence to the feasible region;

[0188] Repair strategy: For slightly infeasible solutions, heuristic methods such as time shifting and resource reallocation are used for repair.

[0189] S43, an adaptive strategy based on the population convergence state dynamically adjusts the crossover probability and mutation probability;

[0190] To address the shortcomings of fixed parameters in traditional genetic algorithms, an adaptive parameter adjustment strategy based on the population convergence state is proposed:

[0191] The first step, convergence state measurement: calculate the standard deviation of the current population fitness. Standard deviation of fitness from the initial population By comparison, the convergence index δ is obtained, δ= / ;

[0192] The second step is to calculate the correlation probability of the upper-level process sequence encoding: the first crossover probability. and the first mutation probability Adjust adaptively according to the following formula;

[0193] ;

[0194] ;

[0195] in, This is the preset lower bound for the first crossover probability. The upper limit of the first crossover probability is preset. This is the preset lower limit of the first mutation probability. This is the preset upper limit of the first mutation probability. This is the adjustment coefficient for the first crossover probability. This is the adjustment coefficient for the first mutation probability;

[0196] Step 3: Population diversity assessment: Calculate the standard deviation of the current population fitness in each generation of evolution. , Quantitatively characterize the degree of population dispersion;

[0197] Step 4: Calculate the relevant probabilities of lower-level resource allocation codes and dynamically adjust the probabilities: based on The size of the second crossover probability is dynamically adjusted. Second mutation probability ;

[0198] Dynamic adjustment is performed using a nonlinear dynamic adjustment formula, which is as follows:

[0199] ;

[0200] ;

[0201] in, This is the base value for the second crossover probability. This is the base value for the second mutation probability. is the preset adaptive adjustment coefficient, and e is the base of the natural logarithm.

[0202] S44, a local optimization operator that reduces job switching;

[0203] First, considering the need for collaborative work among operators, a job switching optimization operator is designed:

[0204] when At the same time, for each operator group, calculate the number of job changes in its process sequence. ;choose The largest group is selected as the optimization target; all processes responsible for the optimization target are extracted and recorded as a process sequence list. , , For list The m-th process in the list, where m is the number of steps in the list. The length of , where Includes process type, start time, and product information; under the premise of meeting process sequence constraints, processes of the same type are arranged according to the principle of adjacent processes. The sorting algorithm, with the core objective of minimizing the number of type switching between adjacent processes, constructs an objective function to group processes of the same type together, thereby reducing frequent switching between different types of processes on the production line. The objective function is:

[0205] ;

[0206] in, To minimize the function, Here, m represents the convergence index, and m is a list. length, Indicate extraction The process type of the j-th process. Indicate extraction The process type of the (j+1)th process; when hour Take 1, otherwise Set to 0;

[0207] Then, maintaining the dependencies between processes, recalculate the start and finish times of each process after the reorganization, verify whether the new solution meets all constraints, and replace the original solution if it does.

[0208] This invention also includes: implementing the algorithm using MATLAB software and solving it in code to output a resource load map, the specific process of which is as follows;

[0209] Algorithm Implementation and Solution Steps: The improved adaptive genetic algorithm is implemented using the MATLAB programming platform. The objective optimization function constructed in step S3 and the constraints established in step S2 are converted into a computable mathematical model. Algorithm running parameters are set, including population size, maximum number of iterations, mutation probability range, and local optimization probability. Iterative calculations of the genetic algorithm are executed, evolving generation by generation until the termination condition is met through operations such as population initialization, fitness calculation, crossover mutation, and local optimization. The optimal chromosome code is output, and decoded to generate a complete production process route plan, including product information such as the start time, end time, and resource allocation for each process. The specific values ​​of the algorithm running parameters are shown in Table 1.

[0210] Table 1. Specific values ​​of algorithm running parameters

[0211]

[0212] Visual output generation steps:

[0213] Generate a resource load map, such as Figure 3 , Figure 4 As shown, the horizontal axis represents time, and the vertical axis represents resource load status. Workload curves for the assembly team and inspectors are plotted separately. A value of 1 on the vertical axis indicates a busy state, and 0 indicates an idle state, displaying the busy and idle status of each resource at different time periods. According to... Figure 3 and Figure 4 The load curve results show that the assembly team is in a state of continuous busy operation for a long time, with extremely high resource utilization; the inspectors are in a state of periodic intermittent busy operation; indicating that the optimized production schedule has achieved peak matching between assembly resources and inspection resources, avoiding serious resource conflicts.

[0214] Generate a daily production completion chart, such as Figure 5 As shown, daily output statistics are based on the completion time of the final inspection process, displaying the number of products that complete the final inspection each day. The horizontal axis represents the number of production days, and the vertical axis represents the number of sets produced per day. According to... Figure 5 The bar chart results show that 50 sets of products were completed daily from day 2 to day 6, and 30 sets were completed on day 7, for a total of 280 sets. This indicates that the optimized production scheduling scheme met the requirements of the maximum daily start-up volume and total production task, and verified that the algorithm can effectively balance the production cycle and complete the manufacturing task on time.

[0215] Generate a buffer usage graph, such as Figure 6 As shown, the horizontal axis represents time in hours, and the vertical axis represents buffer occupancy in sets. The solid blue line in the graph represents the actual capacity usage of the buffer, and the dashed red line represents the maximum capacity of 160 sets, reflecting the occupancy of the buffer and the utilization of the products. Figure 6The results show that the buffer occupancy fluctuates periodically, and the peak occupancy does not exceed the capacity limit of 160 sets; this indicates that the optimized production scheduling scheme can effectively control the number of work-in-process and avoid the risk of work-in-process accumulation or buffer overflow.

[0216] The core objective of the process route coordination method for final assembly production is to construct a multi-objective mixed-integer programming model. Considering various constraints, an improved adaptive genetic algorithm is designed to solve this model. The algorithm's double-layer encoding directly maps the process-resource coordination relationship, and the decoding process naturally integrates multiple complex constraints. Simultaneously, the adaptive strategy dynamically adjusts the crossover and mutation probabilities based on the search state, avoiding premature convergence; local optimization operators specifically address key bottlenecks, enhancing the practicality of the solution. The algorithm is implemented using MATLAB software, outputting a resource load map reflecting the load balance of each workstation, providing direct support for production decisions.

[0217] This invention achieves integrated optimization of process routes and production, and can quickly generate feasible, efficient and balanced final assembly process scheduling schemes, significantly improving the synergy and overall efficiency of the production system. It also provides practical solutions for the industry and has important engineering application value and promotion prospects.

Claims

1. A final assembly production process method with coordinated process routes, characterized in that, Includes the following steps: Step S1: Describe the assembly production process problem of process route coordination and set decision variables; Step S2: Perform mathematical model analysis of the process route for the constraints of the final assembly production process problem; the constraints of the final assembly production process problem include buffer capacity constraints, human resource coordination constraints, temperature settling and control component capacity constraints, process route sequence constraints, batch logic constraints, and other basic constraints. Step S3: Analyze the objective to be optimized and construct the corresponding objective optimization function; Step S4: Design an improved adaptive genetic algorithm to solve the objective optimization function; the specific process of step S4 is as follows: S41, establish a two-layer coding mechanism; S42, Fitness Assessment and Constraint Handling; S43, an adaptive strategy based on the population convergence state dynamically adjusts the crossover probability and mutation probability; S44 is a local optimization operator that reduces job switching.

2. The assembly production process method with coordinated process routes according to claim 1, characterized in that, In step S1, the final assembly production process problem is described as follows: Suppose we are manufacturing core components for high-precision products, denoted as h. The total task comprises N sets of products. We divide the total task into K production batches, satisfying the following conditions: Where k = 1, 2, ..., K Let X be the set of indexes of all products included in the k-th production batch; let t be the total task set; let t be the time in days; the processing of each product is decomposed into 10 processes, which include 5 assembly processes and 5 inspection processes. The 5 assembly processes are denoted as A1, A2, A3, A4, and A5, and the set of assembly processes is... The five inspection procedures are denoted as I1, I2, I3, I4, and I5, and the set of inspection procedures is as follows: The operator team is composed of The inspectors assembled Both the operator team and the inspector are considered human resources, denoted as r, where r belongs to... or Buffer zone: Maximum capacity of 160 sets, considered as space resource.

3. The assembly production process method with coordinated process routes according to claim 2, characterized in that, In step S1, the decision variables are set as follows: This indicates the processing status of the j-th process of the i-th product on day t, where i represents the i-th product and j represents the j-th process. If the j-th process of the i-th product begins processing on day t... =1, otherwise =0, ; : Indicates the start time of the j-th process of the i-th product, in days; : Indicates the completion time of the j-th process of the i-th product, in days; : This indicates the resource occupancy status of the j-th process of the i-th product set using human resources r. If the j-th process of the i-th product set uses either the operator team or the inspector, then... =1, otherwise =0, ; : Represents the spatial state of the i-th product on day t. If the i-th product is in the buffer zone on day t, then... =1, otherwise =0, .

4. The assembly production process method with coordinated process routes according to claim 3, characterized in that, In step S2, the constraints of the final assembly production process problem are: Buffer capacity constraint: The number of buffer sets occupied by work-in-process shall not exceed 160 sets; ； in, This indicates that the i-th product is in the buffer zone on day t, meaning that there exists a j-th process that makes... , for The immediate following process; For the i-th set of products The start time of each process; Human resource coordination constraints include operator constraints, inspector constraints, and workshop staffing constraints. Operator constraints: ； in, It is a function of time t, when When, the function value is 1, when When the time is right, the function value is 0; The operator constraint means that each operator group can undertake a maximum of one assembly task at a time. Inspector constraints: ； The inspector constraint means that each inspector can undertake a maximum of one inspection task at a time; Workshop staffing constraints: ； Workshop staffing constraints are used to ensure that the total number of people in the workshop does not exceed 9. Temperature settling and control component capacity constraints include the first assembly process start time constraint and the daily capacity constraint of core control components. For core components requiring constant temperature, the start time of the first assembly process must satisfy the following formula: ； in, The warehousing time of the core component h of the i-th product is in days. This is the first assembly process for core components that require constant temperature. The first assembly process for the core components of the i-th product that require constant temperature. The start time, Three different sub-components representing the core component of Class A. Three different sub-components representing the core components of Class B; Controlling the daily production capacity constraints of core components: ； in, This refers to the set of assembly processes for core components; the daily production capacity constraint for core components indicates that the daily completion quantity of related processes for core components is less than or equal to 40. Process route sequence constraints: A subsequent process can only begin after all preceding processes have been completed. The specific process route description includes process route construction, parallel paths, merging points, branch synchronization, and final inspection. Process route construction: Establish serial and parallel relationships between processes, including two parallel paths A1→I1→A3→I3→A5→I5 and A2→I2→A3→I3→A5→I5, as well as an independent path A4→I4→A5→I5, forming a process network structure; Parallel paths: A1→I1 and A2→I2 can be parallelized; Meeting point: A3 can only begin after I1 and I2 are completed; Branch synchronization: A3→I3 and A4→I4 are performed in parallel. A5 can only begin after both are completed. Final verification: After A5 is completed, proceed to I5; Among them, the process cannot be preempted, each stage includes mandatory inspection, and the inspection is the immediate preceding process of the next assembly; ； in, Let A be the completion time of assembly process A1 for the i-th product. Let I be the start time of the inspection process I1 for the i-th set of products. Let I be the completion time of inspection process I1 for the i-th set of products. Let be the completion time of assembly process A2 for the i-th product. Let I be the start time of inspection process I2 for the i-th set of products. Let I be the completion time of inspection process I2 for the i-th set of products. Let A be the start time of assembly process A3 for the i-th product. Let A be the completion time of assembly process A3 for the i-th product. Let I be the start time of inspection process I3 for the i-th set of products. Let I be the completion time of inspection process I3 for the i-th set of products. Let A5 be the start time of the assembly process A5 for the i-th product. Let A be the completion time of assembly process A4 for the i-th product. Let I be the start time of inspection process I4 for the i-th set of products. Let I be the completion time of inspection process I4 for the i-th set of products. Let A5 be the completion time of the assembly process A5 for the i-th product. Let I be the start time of inspection process I5 for the i-th set of products; Batch logic constraint: Products are produced sequentially, and the production line must be cleared between batches; ； in, To minimize the function, To maximize the function, Let this be the start time of the first process for the i-th product. Let M be the completion time of the final process for the i-th product, and M be the total number of processes in the process route. This is the set of indexes for all products included in the (k+1)th production batch. The batch logic constraint means that the earliest start time of the first process of any product in the (k+1)th production batch must be later than or equal to the latest completion time of the last process of all products in the kth production batch. Other fundamental constraints include process time and exclusive human resources; Process time: , For process Standard working hours; Exclusive access to human resources: Each process requires only one type of human resource.

5. The assembly production process method with coordinated process routes according to claim 4, characterized in that, The specific process of step S3 is as follows: Step S31, Optimization Target Analysis: Based on Step S1 and Step S2, the core optimization target is defined as minimizing the total production cycle; Step S32, Objective Function Construction: Construct the objective optimization function that minimizes the total production cycle. The mathematical expression is: ； in, Indicates the total production cycle. Let I be the completion time of inspection process I5 for the i-th set of products. The objective function is to minimize the total production cycle.

6. The assembly production process method with coordinated process routes according to claim 5, characterized in that, The specific process of step S41 is as follows: First, the upper-level process sequence is encoded: process-based encoding is used, and the chromosome is represented as a first vector. , , Let L be the L-th gene in the first vector G, where L is the first vector. The length of L is equal to the total number of processes. Let y be a binary tuple (y, d), where y is the product number and d is the process type number, and d = 1, 2, ..., 10 correspond to 5 assembly processes and 5 inspection processes respectively; Then, the lower-level resource allocation coding is performed: a human resource-based coding method is adopted, and the chromosome is represented as a second vector. , It is the same length as G; For the Lth gene in the second vector F, This indicates the resource number assigned to the corresponding process. For assembly processes, then ∈ , indicating one of the operator teams, if To inspect the process, then ∈ , indicating one of the inspectors.

7. The assembly production process method with coordinated process routes according to claim 6, characterized in that, The specific process of step S42 is as follows: The fitness is calculated based on the objective function, using the following formula: ； Where λ is the penalty coefficient. For fitness; P is the penalty for constraint violation; The formula for calculating the penalty item P for constraint violation is: ； in, for The weighting coefficients, for The weighting coefficients, Let Q be the weighting factor for Q, where Q represents the violation of other constraints. Let t be the number of product sets that are in the buffer zone and have not been transferred on day t. Let t represent the number of people present in the workshop at the same time on day t. The constraint handling strategy is as follows: Hard constraint handling: Hard constraints such as process sequence and temperature settling are forcibly satisfied during the decoding process; Soft constraint handling: Buffer capacity, workshop staffing, etc. are handled through penalty functions to guide the search towards convergence to the feasible region; Repair strategy: For slightly infeasible solutions, heuristic methods such as time shifting and resource reallocation are used for repair.

8. The assembly production process method with coordinated process routes according to claim 7, characterized in that, The specific process of step S43 is as follows: To address the shortcomings of fixed parameters in traditional genetic algorithms, an adaptive parameter adjustment strategy based on the population convergence state is proposed: The first step, convergence state measurement: calculate the standard deviation of the current population fitness. Standard deviation of fitness from the initial population By comparison, the convergence index δ is obtained, δ= / ; The second step is to calculate the correlation probability of the upper-level process sequence encoding: the first crossover probability. and the first mutation probability Adjust adaptively according to the following formula; ； ； in, This is the preset lower bound for the first crossover probability. The upper limit of the first crossover probability is preset. This is the preset lower limit of the first mutation probability. This is the preset upper limit of the first mutation probability. This is the adjustment coefficient for the first crossover probability. This is the adjustment coefficient for the first mutation probability; Step 3: Population diversity assessment: Calculate the standard deviation of the current population fitness in each generation of evolution. , Quantitatively characterize the degree of population dispersion; Step 4: Calculate the relevant probabilities of lower-level resource allocation codes and dynamically adjust the probabilities: based on The size of the second crossover probability is dynamically adjusted. Second mutation probability ; Dynamic adjustment is performed using a nonlinear dynamic adjustment formula, which is as follows: ； ； in, This is the base value for the second crossover probability. This is the base value for the second mutation probability. is the preset adaptive adjustment coefficient, and e is the base of the natural logarithm.

9. The assembly production process method with coordinated process routes according to claim 8, characterized in that, The specific process of step S44 is as follows: First, considering the need for collaborative work among operators, a job switching optimization operator is designed: when At the same time, for each operator group, calculate the number of job changes in its process sequence. ;choose The largest group is selected as the optimization target; all processes responsible for the optimization target are extracted and recorded as a process sequence list. , , For list The m-th process in the list, where m is the number of steps in the list. The length of , where Includes process type, start time, and product information; under the premise of meeting process sequence constraints, processes of the same type are arranged according to the principle of adjacent processes. The sorting algorithm, with the core objective of minimizing the number of type switching between adjacent processes, constructs an objective function to group processes of the same type together, thereby reducing frequent switching between different types of processes on the production line. The objective function is: ； in, To minimize the function, Here, m represents the convergence index, and m is a list. length, Indicate extraction The process type of the j-th process. Indicate extraction The process type of the (j+1)th process; when hour Take 1, otherwise Set to 0; Then, maintaining the dependencies between processes, recalculate the start and finish times of each process after the reorganization, verify whether the new solution meets all constraints, and replace the original solution if it does.

Images