Heavy industry product assembly scheduling optimization method considering pre-process

Abstract

The invention provides a pre-process-considered heavy industry product assembly scheduling optimization method, which comprises the following steps of: constructing a scheduling model in which a processing stage and an assembly stage are coordinated, part processing completion is taken as a pre-condition for starting a corresponding assembly body, the assembly operation is executed by a worker team and group, and the scheduling model comprises a neat set constraint and a team and group configuration constraint; the neat set constraint ensures that any assembly starts to be assembled only after all required parts are machined; setting a multi-objective optimization function including the minimum maximum completion time, the minimum machining energy consumption and the minimum chemical worker team process time balance rate; solving the scheduling model by adopting an improved non-dominated sorting genetic algorithm, wherein the improvement comprises an elite pool strategy, a virtual target vector optimization strategy and a differential evolution operator introduction strategy; and outputting an optimal scheduling scheme including a part processing sequence, machine distribution, an assembly body assembly sequence and team personnel configuration.

Description

Technical Field

[0001] This invention belongs to the technical fields of intelligent manufacturing and production scheduling optimization, and specifically relates to a method for optimizing the assembly scheduling of heavy industrial products that takes into account the preceding processes. Background Technology

[0002] In the field of modern equipment manufacturing, heavy industrial products (such as engineering machinery, heavy machine tools, large complete sets of equipment, special heavy structural components, etc.) are the core equipment supporting industrial production and infrastructure construction. These products share several key manufacturing characteristics: First, they are large and complex, with individual units often weighing tens to hundreds of tons. This necessitates multi-stage, multi-disciplinary collaborative manufacturing, with each stage tightly integrated; delays in any stage can impact the overall progress. Second, the manufacturing cycle is long, often taking weeks to months from component processing to final assembly. The required components are diverse (including structural parts, transmission components, and control system components), with significantly different processing procedures, precision requirements, and technological demands. Third, assembly is primarily done manually, limited by product structure and assembly space, making automated assembly lines difficult. It often requires collaborative work in teams, where the skill level and teamwork directly affect efficiency and quality. Fourth, the assembly process demands extremely high component completeness; any delayed or substandard component can halt the entire assembly process, triggering a chain reaction.

[0003] Current research on production scheduling in assembly workshops mainly focuses on automated production lines and machine-dominated assembly scenarios (such as automotive parts assembly and electronic equipment assembly). The core of scheduling optimization revolves around "process allocation for single / parallel machines" and "production line cycle time balance," which does not fully adapt to the assembly characteristics and complex requirements of heavy industrial products. Specifically, existing technologies have the following key shortcomings: First, they fail to cover the collaborative scheduling needs of the "processing-assembly" two-stage process for heavy industrial products. In the manufacturing process of heavy industrial products, the sequence of processes, machine selection, and processing efficiency in the processing stage directly determine the material availability, assembly sequence rationality, and overall progress in the subsequent assembly stage. However, existing scheduling studies often focus on a single stage in isolation (either only studying processing scheduling to pursue equipment utilization or only optimizing assembly scheduling to shorten the assembly cycle), failing to establish a collaborative linkage logic between the two stages. This easily leads to a disconnect between the processing and assembly stages, resulting in situations where some parts are completed ahead of schedule and accumulate, while others are delayed, affecting assembly. Second, assembly stage scheduling neglects the core needs of manual collaboration and team management in heavy industrial products. The assembly of heavy industrial products relies on team collaboration, and different teams differ in skill proficiency, personnel configuration, and preferred processes. Furthermore, the assembly tasks... The number of processes and labor intensity vary, requiring a balance of workload among work teams to avoid overload or idleness. However, existing assembly scheduling research often uses "single workstation, single operator" as the scheduling unit, without considering the logic of work team allocation and skill matching, nor establishing an optimization mechanism for balancing work team workload. This results in scheduling schemes that cannot adapt to actual manual assembly scenarios, leading to serious resource waste and potentially affecting assembly quality due to uneven work team loads. Existing scheduling methods rely on manual experience and lack systematic optimization. Currently, in the industry, the processing and assembly scheduling of heavy industrial products is mostly planned by functional department personnel based on past project experience. It is difficult to comprehensively consider multi-dimensional factors such as the status of processing equipment, fluctuations in parts processing cycles, and differences in work team skills. Furthermore, the flexibility to deal with external interference is insufficient, easily leading to a disconnect between scheduling schemes and actual production, making it difficult to achieve efficient resource utilization and maximize production efficiency. Summary of the Invention

[0004] To address the shortcomings and deficiencies of existing technologies, this invention provides a method and system for optimizing the assembly scheduling of heavy industrial products, considering pre-processing steps. This aims to solve problems such as the disconnect between the processing and assembly stages in heavy industrial product manufacturing, uneven resource allocation among manual assembly teams, and the lack of systematic optimization solutions. The core of this invention lies in constructing a scheduling model that coordinates the processing and assembly stages. This model uses the completion of component processing as a prerequisite for starting the corresponding assembly. It introduces assembly completeness constraints to ensure that any assembly only begins assembly after all required parts have been processed. Simultaneously, it innovatively defines team configuration constraints, requiring that the number of workers in a team responsible for a single assembly process is not less than the total number of assembly steps required for that assembly, thereby refining the management of human resources in the manual assembly stage. Based on this, this invention sets up a multi-objective optimization function that minimizes the maximum completion time, minimizes machine processing energy consumption, and minimizes the imbalance in worker team process time, comprehensively improving production efficiency, reducing energy consumption, and achieving balanced workload among teams. To efficiently solve the above model, this invention makes targeted improvements to the classic non-dominated sorting genetic algorithm, integrating an elite pool strategy to dynamically retain high-quality solutions, a virtual target vector optimization strategy to guide the search direction, and a differential evolution operator introduction strategy to enhance global search capabilities. The final output includes the optimal scheduling scheme that includes the processing order of parts and machine allocation, the assembly order of assemblies, and the configuration of team personnel, effectively supporting intelligent decision-making and resource optimization in heavy industry product assembly workshops.

[0005] The specific technical solution adopted by this invention to solve its technical problem is as follows:

[0006] An optimization method for assembly scheduling of heavy industrial products that considers preceding processes includes:

[0007] A scheduling model is constructed that coordinates the processing and assembly stages, with the completion of part processing as a prerequisite for the start of the corresponding assembly. The assembly operation is performed by worker teams and includes a completeness constraint and a team configuration constraint. The completeness constraint ensures that any assembly will only start assembly after all required parts have been processed. The team configuration constraint ensures that the number of workers in the team responsible for the process of a single assembly is not less than the total number of assembly steps required for that assembly.

[0008] Set up a multi-objective optimization function that includes minimizing the maximum completion time, minimizing machine processing energy consumption, and minimizing the worker shift process time balance rate;

[0009] An improved non-dominated sorting genetic algorithm is used to solve the scheduling model. The improvements include an elite pool strategy, a virtual target vector optimization strategy, and a differential evolution operator introduction strategy.

[0010] The output includes the optimal scheduling scheme for the processing sequence of parts and machine allocation, assembly sequence of assemblies, and team personnel configuration.

[0011] Furthermore, the associated constraints of the scheduling model also include: each process in the processing stage can only be processed by one machine; the start time of the subsequent process of the same part is not earlier than the completion time of its preceding process; each worker can only belong to one assembly team; the start time of the process is not earlier than its completion time, and both the start time and the completion time of the process are non-negative values.

[0012] Furthermore, the specific definition of the multi-objective optimization function is as follows: minimize the maximum completion time, which is the time when the assembly task is completed latest among all assemblies; minimize the machine processing energy consumption, which is the sum of the working energy consumption and idle energy consumption of all machines; minimize the process time balance rate of worker shifts, which is the maximum absolute deviation between the process completion time of each shift and the average process completion time of all shifts.

[0013] Furthermore, the elite pool strategy includes: setting an iteration number threshold; when the iteration number exceeds the threshold, dynamically adding non-dominated solution individuals generated in each generation to the elite pool; using a fast non-dominated sorting method to retain individuals in the first non-dominated layer of the elite pool; and adopting an adaptive adjustment mechanism for the capacity of the elite pool: when the capacity is insufficient for the population size, supplementing in descending order of crowding distance; and when the capacity exceeds the population size, truncating in descending order of crowding distance.

[0014] Furthermore, the virtual target vector optimization strategy includes: triggering weight vector updates every preset number of generations; generating a virtual target vector based on the target vector of the current elite pool individual and preset adjustment parameters; and recalculating reference points and neighbor relationships based on the virtual target vector to guide the algorithm to converge toward the optimal solution.

[0015] Furthermore, the differential evolution operator introduction strategy includes: triggering the differential evolution operator with a preset probability when the number of iterations reaches a preset proportion of the total number of iterations; randomly selecting two individuals from the elite pool as parents, and generating new population individuals through differential mutation and crossover operations to enhance the global search capability of the algorithm.

[0016] Furthermore, before using the improved non-dominated sorting genetic algorithm to solve the problem, an initialization strategy is also included, specifically: from the set of available machines for each process of the workpiece, select the currently idle machine with the shortest processing time as the processing machine for that process; determine the assembly sequence of the assembly based on the requirement of completeness of parts; allocate the number of workers to each assembly task according to the weight of the process time of each assembly task in the total assembly process time, and allocate workers to each shift in combination with the worker process time data, so that the process time of each shift is similar.

[0017] Furthermore, the improved non-dominated sorting genetic algorithm uses a three-layer coding structure when generating the initial population. Specifically, the first layer is the process and assembly sequence code, where the value of the processing stage represents the corresponding workpiece number, the position order represents the process sequence, and the value of the assembly stage represents the sub-assembly number; the second layer is the machine and worker allocation code, where the value of the machine allocation segment represents the processing machine number of the corresponding workpiece and process, and the value of the shift code segment represents the number of workers in the corresponding assembly shift; the third layer is the worker composition code within the shift, where the value represents the worker number of the corresponding assembly shift.

[0018] Furthermore, the improved non-dominated sorting genetic algorithm for solving the scheduling model includes: performing non-dominated sorting and crowding distance calculation on the initial population; performing selection, crossover, and mutation operations, integrating the elite pool strategy, virtual target vector optimization strategy, and differential evolution operator to update the population; iterating until the algorithm converges or reaches the preset number of iterations, and outputting the Pareto optimal solution set, wherein the optimal scheduling scheme is selected from the Pareto optimal solution set.

[0019] And, a heavy-duty product assembly scheduling optimization system considering upstream processes, comprising:

[0020] The model building module is used to build a scheduling model for the coordination of the processing stage and the assembly stage. The model takes the completion of part processing as a prerequisite for the start of the corresponding assembly. The assembly operation is performed by the worker team. The model includes a kitting constraint and a team configuration constraint: the kitting constraint ensures that any assembly will only start assembly after all the required parts have been processed. The team configuration constraint is that the number of workers in the team responsible for the process of a single assembly is not less than the total number of assembly operations required for that assembly.

[0021] The optimization function setting module is used to set multi-objective optimization functions, including minimizing the maximum completion time, minimizing machine processing energy consumption, and minimizing the worker shift process time balance rate.

[0022] The algorithm solution module is used to solve the scheduling model using an improved non-dominated sorting genetic algorithm. The improvements include an elite pool strategy, a virtual target vector optimization strategy, and a differential evolution operator introduction strategy.

[0023] The solution output module is used to output the optimal scheduling solution, including the processing sequence of parts and machine allocation, assembly sequence of assemblies and team personnel configuration.

[0024] And a computer device including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the method described above.

[0025] A non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described above.

[0026] Compared to existing technologies, this invention and its preferred embodiments offer at least the following advantages: By constructing a scheduling model that deeply coordinates the processing and assembly stages, and introducing a kitting constraint that requires the completion of component processing as a prerequisite for assembly commencement, the invention effectively solves the material waiting and assembly stall problems caused by the disconnect between the two stages in traditional scheduling methods, thereby significantly improving the continuity and efficiency of the overall production process. Through innovatively defining team configuration constraints and setting a multi-objective optimization function centered on minimizing the team process time imbalance rate, this invention enables refined allocation and balanced scheduling of assembly team human resources, avoiding the resource waste phenomenon common in manual assembly scenarios where some teams are overloaded and others are idle, thus improving the efficiency of human resource utilization. Furthermore, by making targeted improvements to the non-dominated sorting genetic algorithm, integrating the elite pool strategy, virtual target vector guidance, and differential evolution operator, the convergence and solution quality of the algorithm in solving multi-objective optimization problems under complex constraints are enhanced. This enables scheduling schemes to be generated quickly and simultaneously take into account multiple indicators such as completion time, energy consumption, and manpower balance, providing highly systematic and operable optimization decision support for heavy industry product assembly workshops. Attached Figure Description

[0027] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0028] Figure 1 This is a flowchart illustrating the implementation of the optimization solution method in an embodiment of the present invention.

[0029] Figure 2 This is a schematic diagram of the chromosome coding structure in an embodiment of the present invention. Detailed Implementation

[0030] To make the features and advantages of the present invention more apparent and understandable, specific embodiments are described below in detail:

[0031] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0032] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0033] To address the shortcomings of existing technologies, the present invention aims to provide a method for optimizing the assembly scheduling of heavy industrial products by considering preceding processes. This method offers a systematic scheduling scheme to achieve efficient resource utilization in the processing-assembly process, shorten production cycles, and reduce costs. The implementation of this scheme includes: describing the optimization model for the assembly scheduling of heavy industrial products considering preceding processes; defining parameters and decision variables; defining the correlation constraints between preceding processes and product assembly; constructing an objective function and integrating it to obtain the problem model; solving the problem model using an improved NSGA algorithm; outputting the optimal scheduling scheme, obtaining the optimal processing sequence and machine selection for parts, the optimal assembly sequence, and the optimal arrangement of assembly team personnel. This invention considers the joint constraints of the processing and assembly stages to optimize the assembly scheduling of heavy industrial products and improve workshop production efficiency.

[0034] The construction process of the embodiment of the present invention can be described in the following steps:

[0035] Step S1: Construct a heavy product assembly scheduling model that considers preceding processes, including a description of the heavy product assembly scheduling problem considering preceding processes using symbols and the definition of decision variables;

[0036] Step S2: Define the optimization objectives for constraints, maximum completion time, machine processing energy consumption, and worker shift process time balance rate using symbols;

[0037] Step S3: Design the initialization strategy for machine selection of the workpiece and worker selection of the assembly team;

[0038] Step S4: Design the differential evolution (DE) operator and virtual target vector (VOV) optimization strategy based on the elite pool strategy;

[0039] Step S5: Determine the optimal solution for a heavy product assembly scheduling problem model that considers upstream processes by using the NSGA algorithm and combining it with the above decision scheme, and output the scheduling results.

[0040] In step S1 of this embodiment, a problem model for the scheduling of heavy product assembly considering preceding processes is constructed: the scheduling problem of the heavy product assembly workshop is described as follows: the heavy product production process includes The workpiece processing stage and this The workpiece was assembled manually. The assembly stage of an assembly. During the manufacturing stage, there are... Each workpiece passed through Each process involves multiple steps, and each step has a corresponding set of machines. During the assembly stage, these workpieces undergo assembly processes to obtain... Each assembly All required parts must be machined during the manufacturing phase before assembly can begin; the assembly process has a completeness requirement. Ultimately, [the assembly process is completed by...]. The assembly of individual components yields the marine heavy industry products manufactured in the workshop. During the assembly process, [the following steps are taken]: Composed of workers Each assembly team operates on a manual basis, with different workers... Different assembly teams have varying operational capabilities (assembly time), resulting in different operational capabilities. During the machining phase, the machining sequence of parts and machine allocation are determined by the process. During the assembly phase, workers are assigned to form different assembly teams and team groups. The assembly sequence of each part is determined to optimize multiple performance indicators during the assembly of heavy industrial products. The parameters defined include:

[0041] Part index;

[0042] Process index;

[0043] Worker index;

[0044] Machine index;

[0045] Assembly index;

[0046] Work group index;

[0047] , number of parts;

[0048] Number of workers;

[0049] Number of processes;

[0050] Number of machines;

[0051] Number of assemblies;

[0052] Number of work teams;

[0053] , ;

[0054] Process set };

[0055] Workers' Assembly ;

[0056] machine set {1,2…s};

[0057] Assembly set };

[0058] Class group set = { ;

[0059] Other parameters and variables in the model are represented by the following mathematical symbols:

[0060] ,Component Process set;

[0061] ,Component One process;

[0062] Assembly ;

[0063] Assembly process ;

[0064] , Optional machine set;

[0065] The working time of machine m;

[0066] The idle time of machine m;

[0067] , Working hours;

[0068] work group For assemblies process Working hours;

[0069] ,team For assemblies ;

[0070] Component No. The start time of each process;

[0071] ,Component No. The end time of each process step;

[0072] , The time when all processes are completed;

[0073] Assembly process End time;

[0074] Assembly Task completion time;

[0075] The energy consumption of machine m;

[0076] The no-load energy consumption of machine m;

[0077] The process completion time for team g;

[0078] A maximum value;

[0079] Decision variables, if the workpiece In the process Select machine If the value is 1, then the decision variable is 1; otherwise, it is 0.

[0080] Decision variables, if workers Selected to join the assembly team Execute assembly If the assembly task is successful, the decision variable is 1; otherwise, it is 0.

[0081] Decision variables, if process In the process Previously in the machine If the processing is performed, the decision variable is 1; otherwise, it is 0.

[0082] In step S2, the optimization objectives include maximum completion time, machine processing energy consumption, and worker shift process time balance rate.

[0083] The constraints are specifically expressed as follows:

[0084] (1)

[0085] (2)

[0086] (3)

[0087] , (4)

[0088] , (5)

[0089] , (6)

[0090] , (7)

[0091] , (8)

[0092] Equation (1) indicates that each process of a component can only occupy one machine platform; Equation (2) indicates that the start time of any process is earlier than or equal to its completion time; Equation (3) indicates that for the same part, the start time of the subsequent process must be later than or equal to the completion time of the preceding process; Equations (4) and (5) indicate that different processes can use the same machine platform but cannot start at the same time; Equation (6) indicates that the start time and end time of a component must be positive; Equation (7) indicates that a worker can only be assigned to one work group; Equation (8) indicates that the number of workers in the work group responsible for the process of an assembly must be greater than or equal to the number of processes in the assembly task.

[0093] The maximum completion time formula (9), the machine processing energy consumption formula (10), and the worker team process time balance rate formula (11) correspond to three optimization objectives, which can be described as:

[0094] )= (9)

[0095] = (10)

[0096] = (11).

[0097] Step S3 specifically includes the following steps:

[0098] Step S31: Determine the set of available machines for the workpiece, and select the machine that is currently idle and has a short processing time from the set of available machines for the process as the processing machine;

[0099] Step S32: Determine the assembly sequence according to the "completeness" requirement; sort the assembly tasks in order of completing the processing of all required parts.

[0100] Step S33: Determine the number of workers for each assembly task in a weighted manner based on the required process time of the assembly task; for example: if the total assembly task requires 100 hours and the total number of workers is 15, then a task that requires 20 hours will be allocated approximately 3 workers.

[0101] Step S34: Based on the workers' processing time and the worker number constraints obtained in the above steps, workers are assigned to the assembly teams in sequence according to the assembly order.

[0102] Step S4 specifically includes the following steps:

[0103] Step S41: Set the elite pool generation threshold and selection mechanism. When the number of iterations exceeds the threshold, the non-dominated solution individuals generated in each generation are dynamically added to the elite pool EP. The first non-dominated layer individuals in the elite pool are retained by the fast non-dominated sorting method. The elite pool capacity has an adaptive adjustment mechanism. When the elite pool capacity is less than the population size, individuals are repeatedly selected in descending order of crowding distance until the capacity is sufficient. When the elite pool capacity exceeds the population size, individuals are truncated in descending order of crowding distance.

[0104] Step S42: Trigger the update of the weight vector every G generations. Generate a virtual target vector VOV based on the target vector of the current elite pool EP individuals and the adjustment parameter θ. Then recalculate the reference point and neighbor relationship based on the VOV vector.

[0105] Step S43: When the number of iterations reaches the later stage, the differential evolution (DE) operator is introduced to randomly select two individuals from the elite pool EP as the parents of differential evolution.

[0106] Based on the above decision-making scheme, the algorithm adopts a three-layer encoding scheme when generating the initial population, such as... Figure 2 As shown:

[0107] The first layer consists of process and assembly sequence codes, used to confirm the processing order of each process and the order of assembly tasks. The numerical value of the processing stage represents the corresponding workpiece number, and the positional order represents which process it is. After all workpieces are sorted, there is the sequence code of the assembly task, whose numerical value represents the sub-assembly number.

[0108] The second layer is the machine and worker number allocation code. The value of the machine allocation segment represents the machine number used for the corresponding process of the corresponding workpiece. After the machine code is the team code, the value of which represents the number of workers in the team.

[0109] The third layer represents the composition of workers within a work group, with the numerical value representing the worker's number.

[0110] In step S5, a problem model for the assembly scheduling of heavy-duty products considering preceding processes is constructed. When generating the initial population using the NSGA algorithm and the aforementioned decision-making scheme, the above three-layer encoding is employed. The optimal solution for solving the assembly scheduling problem model for heavy-duty products considering preceding processes in this invention is determined using the NSGA algorithm and the aforementioned decision-making scheme. Figure 1 As shown, the specific steps include:

[0111] Step S51: Initialize the initial population according to the encoding method;

[0112] Step S52: Calculate the objective function value of the population and perform non-dominated sorting of the population based on the function value;

[0113] Step S53: Iterate to the set value, use non-dominated sorting, crowding distance, randomization strategy and other methods to obtain elite individuals and store the elite individuals in the elite pool;

[0114] Step S54: Every G generations, use the elite pool and adjustment parameters to calculate the virtual vector VOV to generate new reference points and neighbors, and then generate a new population;

[0115] Step S55: Iterate to the later stage of optimization and use the differential evolution (DE) operator to perform differential optimization on the parent generation in the elite pool to generate new individuals in the population;

[0116] Step S56: Determine whether the number of iterations has reached the set value. If not, return to step S52 and finally output the optimal solution set.

[0117] Compared with the prior art, the present invention has the following beneficial effects:

[0118] 1. This invention considers the joint scheduling of the two stages of heavy industrial product processing and assembly, and solves the scheduling optimization problems of process sequencing, machine selection, assembly sequencing, and team worker selection. With the optimization objectives of maximum completion time, total machine energy consumption, and team worker process time balance rate, it can quickly and effectively obtain a better scheduling scheme and improve production efficiency.

[0119] 2. This invention utilizes an elite pool strategy and introduces differential evolution (DE) operators and VOV weighted guiding vectors as a method for multi-objective decision-making. Combined with the NSGA algorithm, it can be used to solve the shop floor scheduling problem.

[0120] The implementation process of the present invention will be further demonstrated and introduced through a more specific test example:

[0121] Step S1: Construct a heavy product assembly scheduling model that considers preceding processes, including a description of the heavy product assembly scheduling problem that considers preceding processes using symbols and the definition of decision variables.

[0122] The core of this step is to instantiate the mathematical model defined in the specification based on the above application scenario. This embodiment takes a marine platform structural component assembly workshop as the application scenario. This workshop needs to process four types of core components, each component requires three processes, and is finally assembled into two large sub-assemblies (components 1 and 2 form assembly 1, and components 3 and 4 form assembly 2). The component workshop has seven processing machines, with two parallel machines each for process 1 and process 2, and three parallel machines for process 3, as well as seven assembly workers with different skill levels who need to form two shifts to complete two assembly tasks.

[0123] Step S2: Define the optimization objectives for constraints, maximum completion time, machine processing energy consumption, and worker shift process time balance rate using symbols;

[0124] Apply the constraints (1) to (8) and the optimization objectives (9) to (11) of step S2 to this example. For example, constraint (1) ensures that each process of each component can only be processed on one machine in the corresponding machine set; objective (9) aims to minimize the time of the last completion of the three sub-assemblies.

[0125] Step S3: Initialize the strategy;

[0126] Initialize according to the strategy in claim 3:

[0127] Step S31: For each process of each workpiece, select the machine with the shortest processing time from its set of available machines.

[0128] Step S32: Based on the "completeness" requirement, which states that all parts required for an assembly must be completed in the machining stage before assembly can begin, the assembly sequence is determined. In this embodiment, the assembly whose required parts are completed first is assembled first.

[0129] Step S33: Allocate the number of workers proportionally based on the total estimated process time for each assembly task. In this embodiment, the total assembly task is set to require 70 hours, assembly 1 requires 40 hours, assembly 2 requires 30 hours, and the total number of workers is 7. Therefore, assembly 1 requires 4 workers and assembly 2 requires 3 workers.

[0130] Step S34: Based on the assembly sequence and the number of workers determined in step S33, assign the workers to each assembly team in turn, so as to make the process time of each team as similar as possible.

[0131] Step S4: Optimize the execution of the strategy:

[0132] According to claim 4, design an optimization strategy:

[0133] Step S41: In this embodiment, the EP threshold is set to 50 generations. After the algorithm has iterated for more than 50 generations, the non-dominated solutions generated in each generation are added to the EP. The EP capacity is adaptively adjusted according to step S41 in the specification, and is the same as the population size by default (e.g., 100).

[0134] Step S42: Every 20 generations (G=20), generate VOV based on the target vector and parameter θ=0.1 of the individuals in EP, and update the reference point and neighbor relationships;

[0135] Step S43: When the number of iterations exceeds 70% of the total number of iterations, the DE operator is triggered with a probability of 0.3, and two individuals are randomly selected from EP to perform differential mutation and crossover to generate new individuals.

[0136] Step S5: Determine a method for scheduling the assembly of heavy-duty products that considers preceding processes, using the NSGA algorithm in conjunction with the above decision scheme.

[0137] Perform according to the encoding method of claim 5 and step S5 in the specification:

[0138] Step S51: In this embodiment, the workpiece, machine, and worker are encoded according to the encoding method of claim 5, specifically as follows: Figure 2 As shown, a population of 100 individuals is initialized using a three-layer encoding method;

[0139] Steps S52 to S56: Iterative optimization will be performed, calculating the three objective function values ​​for each individual. The algorithm performs non-dominated sorting and crowding calculation, executes selection, crossover, and mutation operations, and incorporates the optimization strategy from step S4. It stops after 500 iterations and outputs the Pareto optimal solution set.

[0140] Decision-makers can select a final scheduling scheme from the Pareto solution set based on actual preferences. This scheme specifies the processing sequence and machine allocation for each part, the assembly sequence for each assembly, and the composition of workers in each assembly team.

[0141] Based on the same inventive concept, this invention also provides a computer device, comprising: one or more processors, and a memory for storing one or more computer programs; the programs include program instructions, and the processor executes the program instructions stored in the memory. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, used to implement one or more instructions, specifically for loading and executing one or more instructions stored in a computer storage medium to implement the above-described method.

[0142] It should be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, performs the above-described method. This storage medium can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0143] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0144] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

[0145] This invention is not limited to the above-described preferred embodiments. Anyone inspired by this invention can derive other forms of optimization methods for the assembly scheduling of heavy industrial products that consider the preceding processes. All equivalent variations and modifications made within the scope of the claims of this invention shall fall within the scope of this invention.

Claims

1. A method for optimizing assembly scheduling of reworked products considering pre-process, characterized in that, The application relates to a scheduling model for a manufacturing and assembly process. The scheduling model is based on a multi-objective optimization function, which includes minimizing the maximum completion time, minimizing the energy consumption of machine processing and minimizing the process time balance rate of the worker team. The scheduling model is solved by an improved non-dominated sorting genetic algorithm, which includes an elite pool strategy, a virtual target vector optimization strategy and a differential evolution operator introduction strategy. The scheduling model includes the processing sequence of parts and machine allocation, the assembly sequence of assembly bodies and the personnel configuration of the team. The multi-objective optimization function includes minimizing the maximum completion time, minimizing the energy consumption of machine processing and minimizing the process time balance rate of the worker team.

2. The method according to claim 1, wherein: The elite pool strategy includes setting an iteration threshold, dynamically adding non-dominated solution individuals generated in each generation to the elite pool when the iteration number exceeds the threshold, reserving the first non-dominated layer individuals in the elite pool by using a fast non-dominated sorting method, and using an adaptive adjustment mechanism to adjust the capacity of the elite pool.

3. The method of claim 1, wherein: The virtual target vector optimization strategy includes triggering weight vector updating every interval of preset generations, generating a virtual target vector based on the target vector of the current elite pool individual and preset adjustment parameters, recalculating the reference point and the neighbor relationship according to the virtual target vector, and guiding the algorithm to converge to the optimal solution.

4. The method of claim 1, wherein: The differential evolution operator introduction strategy includes triggering the differential evolution operator at a preset probability when the iteration number reaches a preset proportion of the total iteration number, randomly selecting two individuals as parents from the elite pool, generating new population individuals through differential mutation and crossover operation, and enhancing the global search ability of the algorithm.

5. The method of claim 1, wherein: ​ 6. The method of claim 1, wherein: ​ 7. The method of claim 1, wherein: Before solving the scheduling model by the improved non-dominated sorting genetic algorithm, an initialization strategy is further included, specifically: from the set of selectable machines of each process of a workpiece, a machine that is currently idle and has a relatively short processing time is selected as the processing machine of the process; an assembly sequence of an assembly body is determined according to the requirement of part completeness; the number of workers for each assembly task is allocated according to the weight of the process time of each assembly task in the total assembly process time, and the workers are allocated to each team in combination with the worker process time data, so that the process time of each team is similar.

8. The method of claim 1, wherein: When generating an initial population by the improved non-dominated sorting genetic algorithm, a three-layer coding structure is adopted, specifically: the first layer is a process and assembly sequence code, the value of the processing stage represents the corresponding workpiece number and the position sequence represents the process sequence, and the value of the assembly stage represents the sub-assembly body number; the second layer is a machine and worker number allocation code, the value of the machine allocation segment represents the processing machine number corresponding to the process of the corresponding workpiece, and the value of the team coding segment represents the number of workers in the corresponding assembly team; the third layer is a worker composition code in the team, and the value represents the worker number in the corresponding assembly team.

9. The method of claim 1, wherein: The process of solving the scheduling model by the improved non-dominated sorting genetic algorithm includes: non-dominated sorting and crowded distance calculation on the initial population; selection, crossover and mutation operations are performed, and an elite pool strategy, a virtual target vector optimization strategy and a differential evolution operator are used to update the population; iteration is performed until the algorithm converges or a preset iteration number is reached, and a Pareto optimal solution set is output, and the optimal scheduling scheme is selected from the Pareto optimal solution set.

10. A system for optimizing assembly scheduling of reworked products taking into account pre-assembly processes, characterized in that, It comprises: a model construction module for constructing a scheduling model coordinated by processing stages and assembly stages, the model taking the completion of workpiece processing as a precondition for starting the corresponding assembly body, the assembly operation being performed by a worker team, and the model including completeness constraints and team configuration constraints: the completeness constraints ensure that any assembly body starts assembly only after all required parts are processed, and the team configuration constraints ensure that the number of workers in a team responsible for the process of a single assembly body is not less than the total number of assembly processes required by the assembly body; an optimization function setting module for setting a multi-objective optimization function, the function including minimizing the maximum completion time, minimizing the machine processing energy consumption and minimizing the worker team process time balance rate; an algorithm solving module for solving the scheduling model by an improved non-dominated sorting genetic algorithm, the improvement including an elite pool strategy, a virtual target vector optimization strategy and a differential evolution operator introduction strategy; a scheme output module for outputting an optimal scheduling scheme, including part processing sequence and machine allocation, assembly body assembly sequence and team personnel configuration.

Images