Ship block construction process-level production scheduling method and system based on particle swarm-genetic hybrid algorithm

Abstract

The invention discloses a ship block construction process-level production scheduling method and system based on a particle swarm-genetic hybrid algorithm. The method comprises the following steps: constructing a standardized process-level process design data model; based on an operation plan decomposition rule, decomposing the segmented production task into a part-level operation production task matched with the process; establishing an optimization target model with the shortest delivery time, the least overdue tasks, the most balanced load and the like, and constraint rules such as lane and line separation and customization station rules, process route and process constraint rules, material and capacity constraint, priority and the like; optimizing the segmented production plan through a particle swarm-genetic hybrid algorithm until an optimization result meeting the requirement is obtained; and according to the feedback of the actual execution condition of the plan, the orders are classified based on a clustering algorithm, and the operation cycle parameters of different processes are dynamically compensated. The method has the advantages of being high in plan fineness, small in operation progress estimation deviation and the like, is high in calculation speed and low in algorithm complexity, fully considers support of process design data of a design end on algorithm implementation, and is very suitable for engineering application.

Description

Technical Field

[0001] The present invention belongs to the field of shipbuilding informationization, and in particular to a process-level production scheduling method and system for ship block construction based on a particle swarm-genetic hybrid algorithm. Background Art

[0002] In recent years, driven by the substantial growth of the shipping market, the number of orders held by domestic ship assembly and construction plants has continued to grow. With the continuous growth of production capacity demand and the continuous improvement of the automation level of workshops related to hull manufacturing, the demand for refined, digital and visual management of workshop production factors and intermediate product manufacturing processes has become increasingly urgent. How to use advanced digital technology to fully stimulate the production efficiency of automated equipment, reasonably arrange production tasks, scientifically guide the coexistence of automation and manual production, and further reduce the waste of ineffective working hours has become an urgent problem to be solved in the industry.

[0003] At present, the workshop-level production plan management in my country's shipbuilding field is mainly based on manual formulation, or the production scheduling is realized by manual dragging for load balancing or material quantity balancing. The plan formulation efficiency is low, the granularity is coarse, and the rationality is insufficient. There is a lack of automatic optimization formulation and scheduling capabilities for multi-station plans at the process level under multiple constraints. Taking the block plan as an example, most domestic shipyards still manage the block plan to the block level, that is, the specific production process in the middle of the block is not planned, and only the start production practice and end time of the block are given, and there is no management to a more detailed process level. In the time dimension, it is generally managed to the week or even the month, that is, the number of blocks produced in the week is used as the planning control object, and the production tasks to be carried out on a specific day are not controlled. Although this control method can meet production needs, the plan is difficult to control, resulting in a generally low plan execution rate and frequent plan changes. At the same time, how to improve the production process and shorten the ineffective connection waiting time between processes is also unknown.

[0004] A few shipyards have introduced fully automatic advanced production scheduling systems, which are mainly dominated by foreign products, such as the advanced production scheduling systems of France's Dassault and Germany's Fuze. Due to their lack of in-depth understanding of the domestic shipbuilding production model, their complex systems usually only have a small part of their functions that can be used in the shipbuilding industry, and they are not well integrated with the production organization model. Whether it is the number of collaborative workstations or the scheduling effect, they are still not good. Summary of the invention

[0005] The purpose of the present invention is to propose a fully automatic advanced scheduling method and system for ship block construction in order to address the problems of poor precision and low accuracy in planning and scheduling of ship block construction workshops, difficulty in achieving fully automatic scheduling, and low plan execution rate. The method aims to utilize the rapid search capability of the particle swarm algorithm and the global search capability of the genetic algorithm to achieve fully automatic, process-level multi-station collaborative automatic planning and scheduling of block construction while taking into account optimization objectives such as material quantity and load balance.

[0006] The technical solution to achieve the purpose of the present invention is: a process-level scheduling method for ship block construction based on a particle swarm-genetic hybrid algorithm, the method comprising:

[0007] Step 1: Build a standardized process design data model at the process level;

[0008] Step 2: Construct the operation plan decomposition rules to decompose the segmented production tasks into component-level production tasks that match the process;

[0009] Step 3: Establish an optimization target model with the shortest delivery time, the least overdue tasks, and the most balanced load, and set up constraints on lanes, lines, and customized processing positions, process routes, and process steps, materials, capacity, and priority.

[0010] Step 4: Based on the optimization target model and constraint rules in step 3, the segmented production plan is optimized by using a particle swarm-genetic hybrid algorithm until an optimization result that meets the preset requirements is obtained;

[0011] Step 5: Based on the clustering algorithm, the orders are classified in combination with the plan execution feedback, and the operation cycle parameters of different processes are dynamically compensated.

[0012] Furthermore, step 1 specifically includes:

[0013] A standardized process design data model is constructed in the ship design PLM or PDM software, which includes a process table, a process equipment relationship table, a process route table, a process detail table under the process, a process connection relationship table, a material table and a material BOM table; the PLM or PDM software automatically generates master data including the above tables and sends it to the production management software.

[0014] Furthermore, step 2 specifically includes:

[0015] Step 2-1, obtain the finished material according to the material type in the material table, that is, the material code and process route ID code of the segmented material;

[0016] Step 2-2, take the finished material as the root node, search the material BOM table and find the corresponding material in the input material, and traverse its corresponding input material ID;

[0017] Step 2-3, store the material input and required quantity of each layer node as child nodes of the tree structure;

[0018] Step 2-4, take the input material of the child node as the output material, repeat steps 2-2 to 2-3 until all materials in the material table are traversed, and store the results in a tree structure;

[0019] Step 2-5, search the material table, match the process route ID and process ID corresponding to each material with each material, and you can get the list of materials and processes that need to be scheduled and cannot be divided under the current segment. Then, combine it with the workstations defined by the standard as elements of the output matrix, and describe the scheduling results from the material perspective and process perspective respectively.

[0020] Furthermore, the constraint rules in step 3 are specifically:

[0021] (1) Lane division and customized workstation constraint rules

[0022] For a certain process, production can only be carried out at a certain station or several parallel stations. This constraint is obtained from the process-equipment relationship table; the Boolean variable a is introduced i,j (τ) represents the corresponding relationship between process i and workstation j within a period of time τ, which has the following relationship:

[0023]

[0024] (2) Process route and process constraint rules

[0025] The processes under the same process route for producing the same intermediate product must satisfy the process sequence constraints given by the process design;

[0026] (3) Material Constraint Rules

[0027] Process i can only be produced at workstation j when materials are available, that is, the previous process must be completed before the next process can be produced; the Boolean variable u is introduced i,j (τ) indicates whether process i can be processed at workstation j within a period of time τ;

[0028]

[0029] Where, t i-1 represents the start processing time of the i-1th process, Δ i-1 represents the processing time of the i-1th process; T i represents the start processing time of the i-th process;

[0030] (4) Capacity Constraint Rules

[0031] The load of a workstation / shift group is 0 within a shift and 1 outside a shift; the load of an unoccupied workstation is 0 and the load of an occupied workstation is 1; the material volume of each workstation and shift per day is fixed, and the part exceeding the production capacity needs to be arranged to a subsequent date or a subsequent shift;

[0032] (5) Priority constraint rules

[0033] When a process is defined with a high priority, it will always be placed at the front of the processes to be scheduled when selecting available workstations and will be given priority in production without being affected by the scheduling algorithm.

[0034] Furthermore, the optimization target model in step 3 is specifically:

[0035] (1) The optimization target model with the shortest delivery time is F1:

[0036] F1=Min(T end )

[0037] Among them, T end Indicates the end time of the segmented task after scheduling;

[0038] (2) The optimization target model with the least overdue tasks is F2:

[0039] F2=MIn(σ)

[0040] Among them, σ represents the variance of workstation utilization;

[0041] (3) The load balance optimization target model is F3:

[0042] F3=Min(count(exceed))

[0043] Among them, count(exceed) represents the number of exceeded processes.

[0044] Furthermore, the calculation formula of the station utilization variance σ is:

[0045]

[0046] Among them, η i Represents the utilization rate of workstation i, n represents the number of workstations, i=1,2,....,n; when the workstation is equipment, η=total production scheduling time of equipment / number of production scheduling days of equipment*daily shift working hours; when the workstation is a manual workstation, η=total production scheduling time of the workstation team / number of production scheduling days*daily shift working hours.

[0047] Furthermore, step 4 specifically includes:

[0048] Step 4-1, initialize the population, and set the population size and the maximum number of iterations; the population is composed of genomes, each genome includes a process segment and a workstation segment, and the workstation segment includes equipment and manual workstations;

[0049] Step 4-2, initialize the PSO particle swarm;

[0050] Step 4-3, traverse all populations, and calculate the fitness of the new population obtained after the genetic algorithm crossover mutation according to the optimization target model in step 3;

[0051] Step 4-4, sort the individuals in the population in descending order according to the fitness value, then select the individual with the highest fitness value as the global optimal position of the PSO algorithm, and then update the speed and position of the particle according to the global optimal position;

[0052] Step 4-5: If the population traversal is completed but the number of iterations has not been reached, the offspring individuals are selected and crossover operations are performed;

[0053] Steps 4-6, perform mutation operations on offspring individuals;

[0054] Step 4-7, repeat the above steps 4-3 to 4-6 until the preset maximum number of iterations is reached, sort the fitness of the population and particle swarm finally obtained, and select the individual with the highest fitness as the optimization result.

[0055] Furthermore, step 5 specifically includes:

[0056] Step 5-1, data preprocessing: convert the collected production scheduling data into a feature vector, where each process is a row, its features are the elements of the vector, and factors related to deviation are selected as features;

[0057] Step 5-2, initialize cluster centers: select the number of clusters K, and randomly select K processes as initial cluster centers;

[0058] Step 5-3, assign samples to the nearest cluster center: For each process, calculate the distance between it and each initial cluster center, and assign it to the cluster represented by the nearest initial cluster center;

[0059] Step 5-4, update the cluster center: for each cluster, calculate the average value of all process characteristic values ​​in the cluster, and use the average value as the new cluster center;

[0060] Step 5-5, repeat the clustering process: repeat steps 5-3 and 5-4 until the change degree of the cluster center is less than the preset standard, or the current iteration number reaches the preset iteration number;

[0061] Step 5-6, obtain clustering results: When the cluster center no longer changes, all processes will be assigned to the final cluster, and the final clustering result, i.e., the production scheduling result, will be obtained;

[0062] Step 5-7: divide all the processes in the production scheduling results into three categories: define the category with more than the preset threshold of overdue processes as category A, define the category with more than the preset threshold of early processes as category B, and the remaining category as category C; and extend the operation cycle of category A processes, shorten the operation cycle of category B processes, and keep the operation cycle of category C processes unchanged;

[0063] Steps 5-8, use the adjusted process operation cycle to schedule production; at the same time, after production, use the new production schedule data as the basis for constructing the clustering model formed by the above clustering process, and update and optimize the clustering model.

[0064] Compared with the prior art, the present invention has the following significant advantages:

[0065] (1) Considering the impact of insufficient support from the design end for the data required for process-level planning and scheduling at the manufacturing end, which causes the inability to implement process-level planning and scheduling, a standardized process-level process design data model is established to provide standardized and complete data support for manufacturing scheduling.

[0066] (2) Constraint rules that are fully consistent with the segmented construction production organization model have been established, and the production scheduling results are practical only under these constraints.

[0067] (3) A segmented construction process-level scheduling algorithm based on a particle swarm-genetic hybrid algorithm was established, which can combine constraint rules with optimization objectives to achieve fully automatic, process-level segmented plan generation.

[0068] (4) The problem of inaccurate planning and scheduling caused by inaccurate statistical data of actual working hours at the ship production site was considered, and collaborative algorithm self-learning rules and models were proposed to dynamically compensate for process cycle parameters, further improve the accuracy of planning and scheduling, and reduce the estimation deviation of work progress.

[0069] The present invention is further described in detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 The present invention is a flow chart of a process-level production scheduling method for ship block construction based on a particle swarm-genetic hybrid algorithm in one embodiment.

[0071] Figure 2 The figure is a flowchart of the segmented operation plan decomposition in one embodiment.

[0072] Figure 3 The figure is a flowchart of segmented process-level production scheduling based on a particle swarm-genetic hybrid algorithm in one embodiment.

[0073] Figure 4 The figure is a flow chart of a collaborative algorithm self-learning model in one embodiment. DETAILED DESCRIPTION

[0074] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0075] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0076] In one embodiment, in combination Figure 1 , a process-level scheduling method for ship block construction based on a particle swarm-genetic hybrid algorithm is provided, the method comprising:

[0077] S1, build a standardized process design data model at the process level;

[0078] S2, construct the operation plan decomposition rules to decompose the segmented production tasks into component-level production tasks that match the process;

[0079] S3, establish an optimization target model with the shortest delivery time, the least overdue tasks, and the most balanced load, and set up constraints on lanes, lines, and customized processing positions, process routes, and process steps, materials, capacity, and priority.

[0080] S4, based on the optimization target model and constraint rules of S3, the segmented production plan is optimized through the particle swarm-genetic hybrid algorithm until the optimization result that meets the preset requirements is obtained;

[0081] S5, based on clustering algorithm, classifies orders in combination with plan execution feedback and dynamically compensates for the cycle parameters of different processes.

[0082] Furthermore, in one embodiment, S1 specifically includes:

[0083] A standardized process design data model is constructed in the ship design PLM or PDM software, which includes a process table, a process equipment relationship table, a process route table, a process detail table, a process connection relationship table, a material table, and a material BOM table; the PLM or PDM software automatically generates master data including the above tables and sends it to the production management software. The specific contents are as follows:

[0084] The process table gives the process ID of each workstation. When producing different intermediate products, the process ID of the same type of process is different and is a unique identifier. At the same time, the parameters such as the operation time, load type, whether it can be interrupted, whether it is a constant time, etc. of the process are given to jointly determine the specific participation mode of the process in the scheduling.

[0085] The process equipment relationship table specifies the production equipment ID required for each process ID. For the workstations in parallel collaborative relationships, each process ID corresponds to all parallel production equipment IDs, indicating that the process can be randomly assigned to any of the equipment during scheduling.

[0086] The process route table specifies all process routes included in the segmented construction. Each intermediate product corresponds to a process route ID, which identifies the processing method of the intermediate product. At the same time, the type of the process route is specified, including linear without overlap, linear with overlap, tree structure, and partial sequence (non-linear). Among them, the branch lines are all linear without overlap, and the main lines are all tree structures, that is, a mixed structure of series and parallel.

[0087] The process details table under the process specifies the number of processes included in each process route.

[0088] The process connection relationship table specifies the process connection sequence under each process, including the series connection relationship and the parallel connection relationship. This table determines the process sequence that a certain intermediate product goes through.

[0089] The material table specifies all material IDs, material descriptions, and material types involved in a segment. Material types include raw materials, intermediate semi-finished products, and finished products.

[0090] The material BOM gives the output material ID, the corresponding process route, the corresponding feed material ID, and the process number corresponding to the feed material, that is, the process number corresponding to the assembly process in a certain process route.

[0091] PLM or PDM software can automatically generate master data containing the above tables, and send the above data to the production management system from the intermediate database or data interface by calling the interface.

[0092] Further, in one embodiment, in combination Figure 2 , S2 specifically includes:

[0093] S2-1, obtain the finished material according to the material type in the material table, that is, the material code and process route ID code of the segmented material;

[0094] S2-2, take the finished material as the root node, search the material BOM table and find the corresponding material in the input material, and traverse its corresponding input material ID;

[0095] S2-3, storing the input material and required quantity of each node in each layer as child nodes of a tree structure (each node represents both the input material of the upper node and the output material of the next node);

[0096] S2-4, taking the input material of the child node as the output material, repeating steps 2-2 to 2-3 until all the materials in the material table are traversed, and storing the results in a tree structure;

[0097] S2-5, search the material table, match the process route ID and process ID corresponding to each material with each material, and you can get the list of materials and processes that need to be scheduled and cannot be divided under the current segment. Then, combine them with the standard defined workstations as elements of the output matrix, and describe the scheduling results from the material perspective and process perspective respectively. (The number of each workstation, the corresponding process route ID, the process under the process route, and the type of material that can be produced by the workstation are all defined in a standardized format, that is, there is a corresponding relationship between materials and workstations, and there is also a corresponding relationship between process routes and processes and workstations. Among them, the process route and workstation are strictly one-to-one, and the material, process and workstation are many-to-one).

[0098] Furthermore, in one embodiment, each constraint rule in S3 is specifically:

[0099] (1) Lane division and customized workstation constraint rules

[0100] For a certain process, production can only be carried out at a certain station or several parallel stations. This constraint is obtained from the process-equipment relationship table; the Boolean variable α is introduced i,j (τ) represents the corresponding relationship between process i and workstation j within a period of time τ, which has the following relationship:

[0101]

[0102] (2) Process route and process constraint rules

[0103] The processes under the same process route for producing the same intermediate product must satisfy the process sequence constraints given by the process design;

[0104] (3) Material Constraint Rules

[0105] Process i can only be produced at workstation j when materials are available, that is, the previous process must be completed before the next process can be produced; the Boolean variable u is introduced i,j (τ) indicates whether process i can be processed at workstation j within a period of time τ;

[0106]

[0107] Where, t i-1 represents the start processing time of the i-1th process, Δ i-1 represents the processing time of the i-1th process; T i represents the start processing time of the i-th process;

[0108] (4) Capacity Constraint Rules

[0109] The load of a workstation / shift group is 0 within a shift and 1 outside a shift; the load of an unoccupied workstation is 0 and the load of an occupied workstation is 1; the material volume of each workstation and shift per day is fixed, and the part exceeding the production capacity needs to be arranged to a subsequent date or a subsequent shift;

[0110] (5) Priority constraint rules

[0111] When a process is defined with a high priority, it will always be placed at the front of the processes to be scheduled when selecting available workstations and will be given priority in production without being affected by the scheduling algorithm.

[0112] Furthermore, in one embodiment, the optimization target model in S3 is specifically:

[0113] (1) The optimization target model with the shortest delivery time is F1:

[0114] F1=Min(T end )

[0115] Among them, T end Indicates the end time of the segmented task after scheduling;

[0116] (2) The optimization target model with the least overdue tasks is F2:

[0117] F2=Min(σ)

[0118] Among them, σ represents the variance of workstation utilization;

[0119] (3) The load balance optimization target model is F3:

[0120] F3=Min(count(exceed))

[0121] Among them, count(exceed) represents the number of exceeded processes.

[0122] Here, in some embodiments, the calculation formula of the station utilization variance σ is:

[0123]

[0124] Among them, η i Represents the utilization rate of workstation i, n represents the number of workstations, i=1,2,....,n; when the workstation is equipment, η=total production scheduling time of equipment / number of production scheduling days of equipment*daily shift working hours; when the workstation is a manual workstation, η=total production scheduling time of the workstation team / number of production scheduling days*daily shift working hours.

[0125] Further, in one embodiment, in combination Figure 3 , step 4 specifically includes:

[0126] S4-1, initialize the population, and set the population size and the maximum number of iterations; the population is composed of genomes, each genome includes a process segment and a workstation segment, and the workstation segment includes equipment and manual workstations;

[0127] S4-2, initialize the PSO particle group; each particle is regarded as an individual in the genetic algorithm population, that is, a gene combination, and the optimization process is limited by resources such as process, equipment, and team;

[0128] S4-3, traverse all populations and calculate the fitness of the new population obtained after the crossover mutation of the genetic algorithm according to the optimization target model in step 3;

[0129] Taking the shortest completion time objective function as an example, the corresponding fitness calculation method is as follows

[0130]

[0131] The input is the completion time of n processes, and its maximum value is the completion time of the last process.

[0132] Fitness calculation:

[0133]

[0134] S4-4, sort the individuals in the population in descending order according to the fitness value, then select the individual with the highest fitness value as the global optimal position of the PSO algorithm, and then update the speed and position of the particle according to the global optimal position;

[0135] S4-5, if the population traversal is completed but the number of iterations has not been reached, the offspring individuals are selected and crossover operations are performed;

[0136] Preferably, in S4-5, a tournament selection method is used to select parent individuals, and a crossover operation is performed to generate child individuals.

[0137] S4-6, mutation operation is performed on offspring individuals;

[0138] S4-7, repeat the above S4-3 to S4-6 until the preset maximum number of iterations is reached, sort the fitness of the population and particle swarm finally obtained, and select the individual with the highest fitness as the optimization result.

[0139] Further, in one embodiment, in combination Figure 4 , S5 specifically includes:

[0140] S5-1, data preprocessing: convert the collected production scheduling data into a feature vector, where each process is a row, its features are the elements of the vector, and factors related to deviation are selected as features;

[0141] Here, the factors related to the deviation include: order quantity and quantity, ship model, process time, worker skill proficiency, equipment failure and order material quantity.

[0142] S5-2, initializing cluster centers: selecting the number of clusters K, and randomly selecting K processes as initial cluster centers;

[0143] Here, preferably, K=3, that is, 3 processes are randomly selected as the initial cluster centers;

[0144] S5-3, assign samples to the nearest cluster center: for each process, calculate the distance between it and each initial cluster center, and assign it to the cluster represented by the nearest initial cluster center;

[0145] S5-4, update cluster center: for each cluster, calculate the average value of all process characteristic values ​​in the cluster, and use the average value as the new cluster center;

[0146] S5-5, repeat the clustering process: repeat S5-3 and S5-4 until the change degree of the cluster center is less than the preset standard, or the current iteration number reaches the preset iteration number;

[0147] S5-6, obtain clustering results: when the cluster center no longer changes, all processes will be assigned to the final cluster, and the final clustering result, i.e., the production scheduling result, will be obtained;

[0148] S5-7, all the processes in the production scheduling results are divided into three categories: the category with more than the preset threshold number of overdue processes is defined as category A, the category with more than the preset threshold number of early processes is defined as category B, and the remaining category is category C; and the operation cycle of the category A process is extended, the operation cycle of the category B process is shortened, and the operation cycle of the category C process remains unchanged;

[0149] S5-8, use the adjusted process operation cycle to schedule production; at the same time, after production, use the new production schedule data as the basis for constructing the clustering model formed by the above clustering process, and update and optimize the clustering model.

[0150] Step 5 refers to the collaborative algorithm self-learning model, which clusters the processes through a clustering model and performs the compensation processing mentioned in steps 5-7 based on the clustering results. The compensation includes "extending or shortening the process operation cycle."

[0151] In one embodiment, a process-level scheduling system for ship block construction based on a particle swarm-genetic hybrid algorithm is provided, the system comprising:

[0152] The first module is used to build a standardized process design data model at the process level;

[0153] The second module is used to construct the operation plan decomposition rules to decompose the segmented production tasks into component-level production tasks that match the process;

[0154] The third module is used to establish an optimization target model with the shortest delivery time, the least overdue tasks, and the most balanced load, and to set up constraints on lanes, lines, and customized processing stations, process routes, and process steps, materials, capacity, and priority.

[0155] The fourth module is used to optimize the segmented production plan based on the optimization target model and various constraint rules of the third module through a particle swarm-genetic hybrid algorithm until an optimization result that meets the preset requirements is obtained;

[0156] The fifth module is used to classify orders based on clustering algorithms and combined with plan execution feedback, and to dynamically compensate for the operation cycle parameters of different processes.

[0157] For the specific definition of the process-level scheduling system for ship block construction based on the particle swarm-genetic hybrid algorithm, please refer to the definition of the process-level scheduling method for ship block construction based on the particle swarm-genetic hybrid algorithm mentioned above, which will not be repeated here. Each module in the above-mentioned process-level scheduling system for ship block construction based on the particle swarm-genetic hybrid algorithm can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0158] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following is achieved:

[0159] Step 1: Build a standardized process design data model at the process level;

[0160] Step 2: Construct the operation plan decomposition rules to decompose the segmented production tasks into component-level production tasks that match the process;

[0161] Step 3: Establish an optimization target model with the shortest delivery time, the least overdue tasks, and the most balanced load, and set up constraints on lanes, lines, and customized processing positions, process routes, and process steps, materials, capacity, and priority.

[0162] Step 4: Based on the optimization target model and constraint rules in step 3, the segmented production plan is optimized by using a particle swarm-genetic hybrid algorithm until an optimization result that meets the preset requirements is obtained;

[0163] Step 5: Based on the clustering algorithm, the orders are classified in combination with the plan execution feedback, and the operation cycle parameters of different processes are dynamically compensated.

[0164] For the specific limitations of each step, please refer to the limitations of the process-level scheduling method for ship block construction based on the particle swarm-genetic hybrid algorithm mentioned above, which will not be repeated here.

[0165] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the computer program implements:

[0166] Step 1: Build a standardized process design data model at the process level;

[0167] Step 2: Construct the operation plan decomposition rules to decompose the segmented production tasks into component-level production tasks that match the process;

[0168] Step 3: Establish an optimization target model with the shortest delivery time, the least overdue tasks, and the most balanced load, and set up constraints on lanes, lines, and customized processing positions, process routes, and process steps, materials, capacity, and priority.

[0169] Step 4: Based on the optimization target model and constraint rules in step 3, the segmented production plan is optimized by using a particle swarm-genetic hybrid algorithm until an optimization result that meets the preset requirements is obtained;

[0170] Step 5: Based on the clustering algorithm, the orders are classified in combination with the plan execution feedback, and the operation cycle parameters of different processes are dynamically compensated.

[0171] For the specific limitations of each step, please refer to the limitations of the process-level scheduling method for ship block construction based on the particle swarm-genetic hybrid algorithm mentioned above, which will not be repeated here.

[0172] The present invention can realize fully automatic, process-level multi-station planning and scheduling, thereby greatly improving the efficiency and accuracy of plan preparation, reducing the deviation of operation progress estimation, improving the plan execution rate, and shortening the time wasted due to ineffective connection and waiting of processes. In the context of the current domestic shipyard production site remaining unchanged and the urgent need to increase production capacity, how to efficiently arrange production plans to achieve balanced and continuous production has become an urgent problem to be solved, so the present invention has broad application prospects. Through the application of the present invention, the continuity and balance of segmented construction production will be further improved, the ineffective waiting time will be reduced, and the workshop manufacturing cycle will be shortened, which will bring considerable economic benefits to shipbuilding enterprises.

[0173] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. A process-level scheduling method for ship block construction based on a particle swarm-genetic hybrid algorithm, characterized in that: The method comprises: Step 1: Build a standardized process design data model at the process level; Step 2: Construct the operation plan decomposition rules to decompose the segmented production tasks into component-level production tasks that match the process; Step 3: Establish an optimization target model with the shortest delivery time, the least overdue tasks, and the most balanced load, and set up constraints on lanes, lines, and customized processing positions, process routes, and process steps, materials, capacity, and priority. Step 4: Based on the optimization target model and constraint rules in step 3, the segmented production plan is optimized by a particle swarm-genetic hybrid algorithm until an optimization result that meets the preset requirements is obtained; Step 5: Based on the clustering algorithm, the orders are classified in combination with the plan execution feedback, and the operation cycle parameters of different processes are dynamically compensated.

2. The process-level scheduling method for ship block construction based on particle swarm-genetic hybrid algorithm according to claim 1 is characterized in that: Step 1 specifically includes: A standardized process design data model is constructed in the ship design PLM or PDM software, which includes a process table, a process equipment relationship table, a process route table, a process detail table under the process, a process connection relationship table, a material table and a material BOM table; the PLM or PDM software automatically generates master data including the above tables and sends it to the production management software.

3. The process-level production scheduling method for ship block construction based on particle swarm-genetic hybrid algorithm according to claim 2 is characterized in that: Step 2 specifically includes: Step 2-1, obtain the finished material according to the material type in the material table, that is, the material code and process route ID code of the segmented material; Step 2-2, take the finished material as the root node, search the material BOM table and find the corresponding material in the input material, and traverse its corresponding input material ID; Step 2-3, store the material input and required quantity of each layer node as child nodes of the tree structure; Step 2-4, take the input material of the child node as the output material, repeat steps 2-2 to 2-3 until all materials in the material table are traversed, and store the results in a tree structure; Step 2-5, search the material table, match the process route ID and process ID corresponding to each material with each material, and you can get the list of materials and processes that need to be scheduled and cannot be divided under the current segment. Then, combine it with the workstations defined by the standard as elements of the output matrix, and describe the scheduling results from the material perspective and process perspective respectively.

4. The process-level scheduling method for ship block construction based on particle swarm-genetic hybrid algorithm according to claim 3 is characterized in that: The specific constraints in step 3 are: (1) Lane division and customized workstation constraint rules For a certain process, production can only be done at a certain station or several parallel stations. This constraint is obtained from the process-equipment relationship table; The Boolean variable a is introduced i,j (τ) represents the corresponding relationship between process i and workstation j within a period of time τ, which has the following relationship: (2) Process route and process constraint rules The processes under the same process route for producing the same intermediate product must satisfy the process sequence constraints given by the process design; (3) Material Constraint Rules Process i can only be produced at workstation j when materials are available, that is, the previous process must be completed before the next process can be produced; The Boolean variable u is introduced i,j (τ) indicates whether process i can be processed at workstation j within a period of time τ; Where, t i-1 represents the start processing time of the i-1th process, Δ i-1 represents the processing time of the i-1th process; T i represents the start processing time of the i-th process; (4) Capacity Constraint Rules The load of a workstation / shift group is 0 within a shift and 1 outside a shift; the load of an unoccupied workstation is 0 and the load of an occupied workstation is 1; the material volume of each workstation and shift per day is fixed, and the part exceeding the production capacity needs to be arranged to a subsequent date or a subsequent shift; (5) Priority constraint rules When a process is defined with a high priority, it will always be placed at the front of the processes to be scheduled when selecting available workstations and will be given priority in production without being affected by the scheduling algorithm.

5. The process-level production scheduling method for ship block construction based on particle swarm-genetic hybrid algorithm according to claim 4 is characterized in that: The optimization target model in step 3 is specifically: (1) The optimization target model with the shortest delivery time is F1: F1=Min(T end ) Among them, T end Indicates the end time of the segmented task after scheduling; (2) The optimization target model with the least overdue tasks is F2: F2=Min(σ) Among them, σ represents the variance of workstation utilization; (3) The load balance optimization target model is F3: F3=Min(count(exceed)) Among them, count(exceed) represents the number of exceeded processes.

6. The process-level production scheduling method for ship block construction based on particle swarm-genetic hybrid algorithm according to claim 5 is characterized in that: The calculation formula of the station utilization variance σ is: Among them, η i Represents the utilization rate of workstation i, n represents the number of workstations, i=1,2,....,n; when the workstation is equipment, η=total production scheduling time of equipment / number of production scheduling days of equipment*daily shift working hours; when the workstation is a manual workstation, η=total production scheduling time of the workstation team / number of production scheduling days*daily shift working hours.

7. The method for scheduling ship block construction process level based on particle swarm-genetic hybrid algorithm according to claim 1 is characterized in that: Step 4 specifically includes: Step 4-1, initialize the population, and set the population size and the maximum number of iterations; the population is composed of genomes, each genome includes a process segment and a workstation segment, and the workstation segment includes equipment and manual workstations; Step 4-2, initialize the PSO particle swarm; Step 4-3, traverse all populations, and calculate the fitness of the new population obtained after the genetic algorithm crossover mutation according to the optimization target model in step 3; Step 4-4, sort the individuals in the population in descending order according to the fitness value, then select the individual with the highest fitness value as the global optimal position of the PSO algorithm, and then update the speed and position of the particle according to the global optimal position; Step 4-5: If the population traversal is completed but the number of iterations has not been reached, the offspring individuals are selected and crossover operations are performed; Steps 4-6, perform mutation operations on offspring individuals; Step 4-7, repeat the above steps 4-3 to 4-6 until the preset maximum number of iterations is reached, sort the fitness of the population and particle swarm finally obtained, and select the individual with the highest fitness as the optimization result.

8. The process-level scheduling method for ship block construction based on particle swarm-genetic hybrid algorithm according to claim 7 is characterized in that: In steps 4-5, the parent individuals are selected using the tournament selection method, and the offspring individuals are generated through crossover operations.

9. The process-level production scheduling method for ship block construction based on particle swarm-genetic hybrid algorithm according to claim 1 is characterized in that: Step 5 specifically includes: Step 5-1, data preprocessing: convert the collected production scheduling data into a feature vector, where each process is a row, its features are the elements of the vector, and factors related to deviation are selected as features; Step 5-2, initialize cluster centers: select the number of clusters K, and randomly select K processes as initial cluster centers; Step 5-3, assign samples to the nearest cluster center: For each process, calculate the distance between it and each initial cluster center, and assign it to the cluster represented by the nearest initial cluster center; Step 5-4, update the cluster center: for each cluster, calculate the average value of all process characteristic values ​​in the cluster, and use the average value as the new cluster center; Step 5-5, repeat the clustering process: repeat steps 5-3 and 5-4 until the change degree of the cluster center is less than the preset standard, or the current iteration number reaches the preset iteration number; Step 5-6, obtain clustering results: When the cluster center no longer changes, all processes will be assigned to the final cluster, and the final clustering result, i.e., the production scheduling result, will be obtained; Step 5-7: divide all the processes in the production scheduling results into three categories: define the category with more than the preset threshold of overdue processes as category A, define the category with more than the preset threshold of early processes as category B, and the remaining category as category C; and extend the operation cycle of category A, shorten the operation cycle of category B, and keep the operation cycle of category C unchanged; Steps 5-8, use the adjusted process operation cycle to schedule production; at the same time, after production, use the new production schedule data as the basis for constructing the clustering model formed by the above clustering process, and update and optimize the clustering model.

10. The process-level production scheduling method for ship block construction based on particle swarm-genetic hybrid algorithm according to claim 9 is characterized in that: The factors related to deviation described in step 5-1 include: order quantity and quantity, ship model, process time, worker skill level, equipment failure and order material quantity.

Images