Method and device for optimizing production sequence of large-scale orders
By combining adaptive large neighborhood search and genetic algorithms, multi-process parallelization technology is used to optimize the order of large-scale order production, the optimization problems in the existing technology are solved, efficient and flexible production sequence optimization is achieved, and production efficiency is improved.
Patent Information
- Application Number
- CN202410987981.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-23
- Publication Date
- 2024-11-15
AI Technical Summary
In the hybrid production model of multiple varieties and multiple production lines, it is difficult for the existing technology to effectively optimize the order of large-scale orders, resulting in increased manufacturing costs and inefficient production efficiency.
Combining the advantages of local search and global search, the adaptive large neighborhood search algorithm and genetic algorithm are adopted to optimize the order production sequence through multi-process parallelization technology, and the adaptive large neighborhood search algorithm is used to generate preliminary optimization results, and the genetic algorithm is used to perform the Final optimization.
It realizes efficient solution space exploration, improves adaptability and flexibility, improves solution efficiency through parallelization technology, overcomes the limitations of a single algorithm, and can quickly find an excellent order production order.
Smart Images

Figure CN118966633A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method and a device for optimizing the production sequence of large-scale orders, belonging to the technical field of automobile production. Background Art
[0002] With the development of the manufacturing industry, the mixed production model of multiple varieties and multiple production lines has become the mainstream and has reduced manufacturing costs to a certain extent. However, it has always been a difficult point in the manufacturing industry to fully balance production and sales according to market demand, schedule production and plan supply chain accordingly. Summary of the invention
[0003] The purpose of the present invention is to overcome the deficiencies in the prior art and to provide a method and device for optimizing the production sequence of large-scale orders, which combines the advantages of local search and global search, has efficient solution space exploration capability, and has strong adaptability and flexibility.
[0004] To achieve the above object, the present invention is implemented by adopting the following technical solutions: In a first aspect, the present invention provides a method for optimizing the production sequence of large-scale orders, comprising: Obtain order data for production of complete vehicles within a set time range; Apply an adaptive large neighborhood search algorithm to the acquired order data to obtain preliminary optimization results; The genetic algorithm is executed on the preliminary optimization results to obtain the final optimization results.
[0005] Furthermore, the method also includes: preprocessing the acquired order data, including grouping by attributes and determining the number of groups according to the order type and preset rules.
[0006] Furthermore, the step of applying an adaptive large neighborhood search algorithm to the acquired order data to obtain a preliminary optimization result includes: Randomly sort the acquired order data to obtain multiple initial solutions, each of which represents a production sequence; Applying multiple local search algorithms to the initial solution generates candidate solutions; Calculate the objective function value of the candidate solutions and select the best candidate solution as the preliminary optimization result.
[0007] Furthermore, in the process of randomly sorting the acquired order data to obtain multiple initial solutions, the random sorting is layered, first generating the order of the vehicle models and then generating the order of the internal groups of the vehicle models based on the order of the vehicle models.
[0008] Furthermore, when applying multiple local search algorithms to the initial solution to generate candidate solutions, a multi-process program is used for optimization, and multiple local search algorithms are used to generate candidate solutions for each initial solution in each process.
[0009] Furthermore, the performing of a genetic algorithm on the preliminary optimization result to obtain a final optimization result includes: Select individuals with higher fitness as parents and generate offspring through crossover and mutation operations; Calculate the fitness function value of the offspring and perform roulette wheel selection based on the fitness to generate a new generation of population; Repeat the above steps until the preset termination condition is met to obtain the final optimization result as the optimized order production sequence.
[0010] In a second aspect, the present invention provides a device for optimizing the production sequence of large-scale orders, comprising: The data acquisition module is used to obtain the order data of the production of the whole vehicle within the set time range; Preliminary optimization module: Applying adaptive large neighborhood search algorithm to the acquired order data to obtain preliminary optimization results; The final optimization module is used to execute the genetic algorithm on the preliminary optimization results to obtain the final optimization results.
[0011] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the aforementioned methods.
[0012] In a fourth aspect, the present invention provides a computer device, comprising: Memory, for storing computer programs / instructions; A processor is used to execute the computer program / instructions to implement the steps of any of the aforementioned methods.
[0013] In a fifth aspect, the present invention provides a computer program product, comprising a computer program / instruction, which implements the steps of any one of the aforementioned methods when executed by a processor.
[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a method and device for optimizing the production sequence of large-scale orders, which combines the advantages of local search and global search, has efficient solution space exploration capability, strong adaptability and flexibility, and can improve the solution efficiency through parallelization technology, and has certain advantages in solving complex optimization problems. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 is a flow chart of a method for optimizing the production sequence of large-scale orders provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of iterative solution of the objective function provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0016] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.
[0017] Embodiment 1, as Figure 1 As shown, this embodiment introduces a method for optimizing the production sequence of large-scale orders, including: Get the order data of the production vehicle within the set time range Randomly sort the acquired order data to obtain multiple initial solutions, each of which represents a production sequence; Applying multiple local search algorithms to the initial solution generates candidate solutions; Calculate the objective function value of the candidate solution and select the best candidate solution as the preliminary optimization result Select individuals with higher fitness as parents and generate offspring through crossover and mutation operations; Calculate the fitness function value of the offspring and perform roulette wheel selection based on the fitness to generate a new generation of population; Repeat the above steps until the preset termination condition is met to obtain the final optimization result as the optimized order production sequence.
[0018] The method for optimizing the production sequence of large-scale orders provided in this embodiment specifically involves the following steps in its application process: 1. Configuration plan Definition of a batch: For example, a sequence [a, a, b, b, c, c, a, a, b, b] is considered a batch if the attributes before and after are the same. The batch of this sequence contains [[a, a],[b, b] [c, c], [a, a],[b, b]] The model solution mainly includes switching, concentration within a single batch, interval between batches, quantity within a single batch and other solution objectives.
[0019] Table 1 Production scheduling constraints ; The present invention uses 17 attributes of 5 types of indicators that are the focus of GAC Automobile Manufacturing Plant's actual production scheduling as scoring indicators for production scheduling results. The indicator weight depends on the priority of each attribute in the actual production scheduling work. Each row represents a vehicle, and each column is an attribute. According to the requirements of each factory for specific attributes, the vehicle production sequence needs to be optimized to obtain the optimal solution.
[0020] 5 types of constraint names: Switch constraints for 5 attributes: vehicle type (weight: absolute priority), sunroof (weight: 4), color (weight: 2), battery (weight: 0), configuration level (weight: 0); Two attributes constraining the batch-to-batch interval: small color (weight: 1, lower bound: 60), two-color car (weight: 1, lower bound: 60), graphite battery (weight: 1, lower bound: 30); The four attributes of quantity constraints within a single batch: small color (weight: 1, upper limit: 15, lower limit: 30), two-color car (weight: 1, lower limit: 4), large color (weight: 1, upper limit: 15), graphite battery (weight: 1, lower limit: 1).
[0021] Algorithm goals: For switching categories and single product concentration (the smaller the value, the better) Score = (Number of manual switches – Number of algorithm switches) / (Number of manual switches) For example, the number of manual vehicle model switching is 383 times, the number of algorithm switching is 206 times, and the score is 0.46214 Switching (for example, when switching between models, the model numbers of the two vehicles are different, the model switching is +1, and the switching index should be as small as possible) First priority Model switching times score Third priority Sunroof switching times score 4 Color switching score 2 Fourth priority Battery switching times score Configuration level switching times score For interval and quantity scores (larger values are better): Score = (algorithm satisfaction rate – manual satisfaction rate) / (manual satisfaction rate) For example, the small color interval satisfaction rate is 0.398, the algorithm satisfaction rate is 0.8, and the score is 1.01 Interval (interval refers to the number of intervals between two batches of cars with required attributes. For example, if there are 60 or more non-small color cars between the two batches of small color A cars and small color B cars, if the interval is greater than or equal to 60, the constraint is satisfied, and if it is less than 60, the constraint is satisfied. For this type of constraint, its satisfaction rate will be counted) Third priority Small color interval score Two-color car interval score Graphite Battery Interval Score Quantity within a single batch (for example, if the quantity of a batch of small color A is required to be greater than or equal to 15 and less than or equal to 30, the quantity satisfies this batch, and the measurement within a single batch is also considered the satisfaction rate) Third priority Small color single batch quantity score Score of quantity in a batch of two-color cars Large color single batch quantity score Graphite Battery Interval Score The second priority is the minimum value of each score sequence of the third priority, which can effectively ensure the correctness of the search and prevent it from falling into a local minimum.
[0022] Objective function construction: Since vehicle model switching is an absolute priority, battery switching, configuration level switching and small color interval priority are 0, the objective function is constructed as follows: [Model switching times score, Min([Model switching times score, Sunroof switching times score, color switching times score, small color interval score, two-color car interval score, graphite battery interval score, small color single batch quantity score, two-color car single batch quantity score, large color single batch quantity score, graphite battery single batch quantity score ) Sunroof switching times score* 4 + color switching times score* 2 + small color interval score + two-color car interval score + graphite battery interval score + small color single batch quantity score + two-color car single batch quantity score + large color single batch quantity score + graphite battery single batch quantity score, Battery switching times score + Configuration level switching times score] The premise for optimizing the second objective is that the first objective is not reduced.
[0023] The premise for optimizing the third objective is that the scores of the first and second objectives do not decrease.
[0024] The premise for optimizing the fourth objective is that the scores of the first objective, the second objective and the third objective do not decrease.
[0025] 2. Solution Perform data preprocessing phase Since the number of orders per day (the number of vehicles produced) is as long as 1,000, directly executing operations such as 2-opt, 2-h-opt, 3-opt, 4-opt, relocate(1) will result in hundreds of millions of new processing sequences, which will take too long for the computer to solve. In addition, a large number of orders are homogeneous, so we need to group and merge homogeneous orders, that is, orders with exactly the same attributes. The processing sequence only needs to consider the processing sequence of this group, and the order within the group does not affect the objective function. For example, if the length of the sequence [a, a, a, a, a…,a] is 60, we can group it based on 30 and assign it to two orders [a1, …,a30][a31, …, a60]. This can greatly reduce the amount of data and group 1,000 vehicle data into about 100.
[0026] For small colors, this application adopts 15 to 30 as a group, for large colors we adopt 15 to 60 as a group, for two-color cars we adopt 1 to 4 as a group, and for graphite batteries we adopt one as a group. Among them, graphite batteries have the first priority; two-color cars have the second priority; large colors have the third priority; small colors have the fourth priority.
[0027] Take an example to illustrate the specific grouping of small colors, such as [[small color A, graphite battery A],.., [small color A, graphite battery A]], which has a total of 60 rows. Since graphite batteries have a higher priority, each row is a group, such as [[small color A, ordinary battery A],.., [small color A, ordinary battery A]], which has a total of 45 rows. At this time, the corresponding remainders need to be considered, such as the remainder of 45 / 15 is 0; the remainder of 45 / 16 is 12; it is not in the range of [15~30]; it is not within the corresponding single batch quantity requirement..., the remainder of 45 / 30 is 0. Find all divisors with remainders of 0 or remainders that meet the requirements of 15~30, and give priority to larger divisors. When num=45, the grouping is 30.
[0028] Execute ALNS stage 60 initial solutions (initial production sequences) are obtained through random sorting. The random sorting here is hierarchical generation. The order of the vehicle models is first generated, and then the order of the internal groups of the vehicle models is generated based on the order of the vehicle models. For example, if there are vehicle models A1, A2, … A8, 8 groups of sequences are generated and spliced.
[0029] Use the multi-process program to enter the optimization module, where the process is set to 60 (the CPU has 60 cores).
[0030] For each initial solution in each process, 2-opt, 2-h-opt, 3-opt, 4-opt, relocate(1), ..., relocate (m1), Exchange(1), ..., Exchange(m2) algorithms are used to update the route pool (m1, m2 can be any integer value greater than 0). Here, one sequence is generated into K sequences through all the above algorithms.
[0031] The objective function corresponding to the route pool is calculated for each process, and the optimal route is obtained to update the corresponding initial solution.
[0032] GA execution phase Parent selection: Select a portion of individuals as parents based on the fitness function value. The selection is normalized based on the fitness function value to increase the chances of better individuals being selected as parents.
[0033] Generate offspring: Genetic algorithm operations such as crossover and mutation are used to operate on parent individuals to generate a new set of offspring individuals. The crossover operation combines the information of two parent individuals to generate new individuals. The present invention adopts uniform crossover and other methods. The mutation operation makes small-scale random changes to individuals to increase the diversity of the population.
[0034] Calculate the fitness function value of the offspring individual: For the generated offspring individuals, calculate their fitness function values to evaluate the quality of their solutions.
[0035] Select offspring: According to roulette selection, a part of offspring individuals are selected as the next generation population. Better individuals have a higher selection probability to increase their chances of being selected as the next generation.
[0036] Repeat steps 3.2, 3.3, and 3.4 until the termination condition is met: The process of calculating the fitness function value, generating offspring, and selecting offspring is repeated until the termination condition is met. The termination condition can be reaching the maximum number of iterations (no better objective function is found for 5 consecutive times), finding a satisfactory solution, or stopping after a certain time (60s).
[0037] Pre-grouping instructions An example to illustrate the above grouping algorithm: For example, if a group of orders belongs to a small color and the total number is 32, we use 15-30 as the grouping basis and we can get: only 16 and 17 can satisfy the constraint that the remainder also loads 15~30. At this time, we should choose the grouping basis with smaller group numbers and smaller remainders, so the grouping basis selected at this time is 16.
[0038] Table 2 Schematic diagram of grouping algorithm ; like Figure 2 As shown in the figure, all indicators have been improved compared with the original algorithm, and the solution time is short, only 8 minutes.
[0039] The technical effects of this embodiment are as follows: The advantages of two algorithms are comprehensively utilized: ALNS is a local search algorithm that explores the solution space through repeated local searches and random perturbations, and can quickly find the local optimal solution; while GA is a global search algorithm that searches a wider solution space through population crossover and mutation, and can find better solutions in the global scope. Combining the advantages of ALNS and GA, it can search at multiple scales, overcoming the limitations of a single algorithm.
[0040] Efficient solution space exploration: ALNS+GA can find the local optimal solution by using ALNS for local search, and further explore through GA's global search capability. ALNS provides a fast local search method that can find a relatively good solution in a short time. GA's global search capability can help jump out of the local optimal solution and find a better solution.
[0041] Adaptability and flexibility: ALNS+GA combines the characteristics of two algorithms and has strong adaptability and flexibility. The ratio and parameter settings of ALNS and GA can be flexibly selected according to the nature and needs of the problem to obtain better performance and results.
[0042] Parallelization capability: ALNS+GA can improve the efficiency of solving problems through parallelization technology. Both ALNS and GA are parallelizable algorithms, which can accelerate the solving process and improve the convergence speed and effect of the algorithm by executing multiple search processes in parallel or using multiple computing nodes.
[0043] In summary, ALNS+GA combines the advantages of local search and global search, has efficient solution space exploration capabilities, strong adaptability and flexibility, and can improve solution efficiency through parallelization technology. This gives ALNS+GA certain advantages in solving complex optimization problems.
[0044] Embodiment 2: This embodiment provides a device for optimizing the production sequence of large-scale orders, comprising: The data acquisition module is used to obtain the order data of the production of the whole vehicle within the set time range; Preliminary optimization module: Applying adaptive large neighborhood search algorithm to the acquired order data to obtain preliminary optimization results; The final optimization module is used to execute the genetic algorithm on the preliminary optimization results to obtain the final optimization results.
[0045] The specific functional implementation of each of the above modules can be found in the relevant contents of the method in Example 1 and will not be elaborated here.
[0046] Embodiment 3, this embodiment provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of any one of the methods described in Embodiment 1 are implemented.
[0047] Embodiment 4: This embodiment provides a computer device, including: Memory, for storing computer programs / instructions; A processor, configured to execute the computer program / instructions to implement the steps of any one of the methods described in Example 1.
[0048] Embodiment 5: This embodiment provides a computer program product, including a computer program / instruction, which implements the steps of any method described in Embodiment 1 when executed by a processor.
[0049] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
[0050] It should be understood by those skilled in the art that the embodiments of the present disclosure may be provided as methods, systems or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0051] The present disclosure is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram and the combination of processes and / or blocks in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0052] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0053] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.
[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present disclosure rather than to limit its protection scope. Although the present disclosure has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that after reading the present disclosure, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the invention, but these changes, modifications or equivalent substitutions are all within the protection scope of the disclosed claims to be approved.
Claims
1. A method for optimizing the production sequence of large-scale orders, characterized in that include: Obtain order data for vehicle production within a set time range; Apply an adaptive large neighborhood search algorithm to the acquired order data to obtain preliminary optimization results; The genetic algorithm is executed on the preliminary optimization results to obtain the final optimization results.
2. The method for optimizing the production sequence of large-scale orders according to claim 1, characterized in that: The method further includes: pre-processing the acquired order data, including grouping by attributes and determining the number of groups according to order type and preset rules.
3. The method for optimizing the production sequence of large-scale orders according to claim 1, characterized in that: The adaptive large neighborhood search algorithm is applied to the acquired order data to obtain preliminary optimization results, including: Randomly sort the acquired order data to obtain multiple initial solutions, each of which represents a production sequence; Applying multiple local search algorithms to the initial solution to generate candidate solutions; Calculate the objective function value of the candidate solution and select the best candidate solution as the preliminary optimization result.
4. The method for optimizing the production sequence of large-scale orders according to claim 3, characterized in that: In the process of randomly sorting the acquired order data to obtain multiple initial solutions, the random sorting is hierarchical, first generating the order of vehicle models and then generating the order of the internal groups of vehicle models based on the order of vehicle models.
5. The method for optimizing the production sequence of large-scale orders according to claim 3, characterized in that: When applying multiple local search algorithms to the initial solution to generate candidate solutions, a multi-process program is used for optimization, and multiple local search algorithms are used to generate candidate solutions for each initial solution in each process.
6. The method for optimizing the production sequence of large-scale orders according to claim 1, characterized in that: The performing of the genetic algorithm on the preliminary optimization result to obtain the final optimization result includes: Select individuals with higher fitness as parents and generate offspring through crossover and mutation operations; Calculate the fitness function value of the offspring and perform roulette wheel selection based on the fitness to generate a new generation of population; Repeat the above steps until the preset termination conditions are met to obtain the final optimization result, which serves as the optimized order production sequence.
7. A device for optimizing the production sequence of large-scale orders, characterized in that include: The data acquisition module is used to obtain the order data of the production vehicle within the set time range; Preliminary optimization module; Apply an adaptive large neighborhood search algorithm to the acquired order data to obtain preliminary optimization results; The final optimization module is used to execute the genetic algorithm on the preliminary optimization results to obtain the final optimization results.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
9. A computer device, characterized in that: include: Memory, used to store computer programs / instructions; A processor, configured to execute the computer program / instructions to implement the steps of the method according to any one of claims 1 to 6.
10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
Citation Information
Cited By
Typesetting method and system for ensuring consistent falling order of different surfaces of hollow glass
CN120317154A
Order production scheduling method and system, electronic equipment and storage medium
CN121998304A
