Intelligent production process scheduling method and system based on super-heuristic genetic algorithm

Abstract

The application relates to the technical field of production scheduling, and provides a production process intelligent scheduling method and system based on a super-heuristic genetic algorithm, which comprises the following steps: initializing and optimizing a scheduling solution set of a product processing plan through an intelligent scheduling mechanism based on the super-heuristic genetic algorithm, and selecting an optimal solution from the scheduling solution set; wherein the intelligent scheduling mechanism based on the super-heuristic genetic algorithm comprises one heuristic chromosome and one production order chromosome, and the intelligent scheduling mechanism based on the super-heuristic genetic algorithm updates the scheduling solution in the following ways: the crossover and mutation operations between the heuristic chromosomes, and the genetic operations of the production order chromosome according to the heuristic operations in the heuristic chromosomes. The method not only has more outstanding scheduling optimization performance, but also guarantees the wide applicability to different optimization targets in different scheduling scenarios.

Description

Technical Field

[0001] This invention belongs to the field of production scheduling technology, and particularly relates to a method and system for intelligent scheduling of production processes based on a hyperheuristic genetic algorithm. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Zero-wait-time intermittent production is a common assembly line production model used for processes that require uninterrupted, continuous operation between adjacent steps to ensure product quality. It is widely used in industries such as pharmaceuticals, fermentation, and textiles. Scientific and efficient production scheduling is not only crucial for ensuring the stable and orderly operation of zero-wait-time intermittent production, but also a powerful means to improve production efficiency and enhance workshop processing performance.

[0004] Scheduling optimization for zero-wait-interruption production processes aims to optimize key production indicators by assigning a reasonable processing order to products while ensuring uninterrupted continuous operation between adjacent processes. Zero-wait-interruption production process scheduling is a crucial form of flow shop scheduling and has been proven to be an NP-hard problem. With the rise of artificial intelligence, heuristic algorithms have been widely applied to solve related scheduling problems. Metaheuristic algorithms, such as simulated annealing and greedy algorithms, are widely used in scheduling optimization across various scenarios due to their low computational cost and ease of implementation. Population-based heuristic algorithms, such as particle swarm optimization and genetic algorithms, have been further promoted and applied due to their wide state space search range, diverse population characteristics, and low risk of getting trapped in local optima, making outstanding contributions, particularly in the scheduling research of zero-wait-interruption production processes.

[0005] While the aforementioned heuristic optimization algorithms have all demonstrated outstanding performance in the scheduling field, any high-performance scheduling optimization based on these algorithms requires improvements to the algorithms based on professional knowledge. Improvements to heuristic scheduling algorithms are often only targeted at specific optimization scenarios and objectives, and heuristic algorithms with broad applicability are often difficult to find. Summary of the Invention

[0006] To address the technical problems mentioned above, this invention provides an intelligent scheduling method and system for production processes based on a hyperheuristic genetic algorithm. This method employs a process of encoding and optimizing low-level heuristic operation mechanisms in high-level heuristic chromosomes, and then using these low-level heuristic operations to evolve scheduling solutions. This achieves self-learning evolution of the genetic operation mechanism, demonstrating significantly superior scheduling optimization performance and ensuring broad applicability to different optimization objectives in various scheduling scenarios.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] The first aspect of this invention provides an intelligent scheduling method for a production process based on a hyperheuristic genetic algorithm, comprising:

[0009] The quantity of products to be processed and the processing information of each product are obtained. The scheduling solution set of the product processing plan is initialized and optimized through an intelligent scheduling mechanism based on a hyperheuristic genetic algorithm.

[0010] The optimal solution is selected from the set of scheduling solutions, and the processing plan represented by the optimal solution is put into the production process of the product.

[0011] Among them, the intelligent scheduling mechanism based on the hyperheuristic genetic algorithm combines a heuristic chromosome and a production order chromosome to form a scheduling solution. A heuristic chromosome consists of several heuristic operations, and a production order chromosome is the full permutation of the codes of all products to be processed. The intelligent scheduling mechanism based on the hyperheuristic genetic algorithm updates the scheduling solution by means of crossover and mutation operations between heuristic chromosomes, as well as genetic operations performed by the production order chromosome based on the heuristic operations in the heuristic chromosome.

[0012] Furthermore, the intelligent scheduling mechanism based on the hyperheuristic genetic algorithm selects several scheduling solutions with the highest comprehensive advantage from the combined parent-child population in each iteration as dominant individuals, forming a dominant solution set, and uses the selected dominant solution set as the parent population for the next round of genetic operations.

[0013] Furthermore, for multi-objective optimization scheduling, the comprehensive dominance is a weighted sum of the dominance and the neighborhood solution distribution density.

[0014] Furthermore, the heuristic operation includes one or more of the following: random swap, adjacent swap, pre-operation, post-operation, reverse operation, and truncation operation.

[0015] Furthermore, for single-objective optimization scheduling, the optimization objective of the intelligent scheduling mechanism based on the hyperheuristic genetic algorithm is the maximum production completion time, the total processing time, or the maximum delay time.

[0016] Alternatively, for multi-objective optimization scheduling, the optimization objectives of the intelligent scheduling mechanism based on the hyperheuristic genetic algorithm are the maximum production completion time, the total processing time, and the maximum delay time.

[0017] Furthermore, the processing information of the product includes: the processing steps and processing time performed on each device.

[0018] Furthermore, for multi-objective optimization scheduling, the optimal solution is selected from the scheduling solution set based on the weighted objective function value;

[0019] The weighted objective function value is:

[0020]

[0021] Where, ω g For the g-th target f g The weight.

[0022] A second aspect of the present invention provides an intelligent scheduling system for a production process based on a hyperheuristic genetic algorithm, comprising:

[0023] The processing plan optimization module is configured to: obtain the quantity of products to be processed and the processing information of each product, and initialize and optimize the scheduling solution set of the product processing plan through an intelligent scheduling mechanism based on a hyperheuristic genetic algorithm;

[0024] The optimal solution selection module is configured to select the optimal solution from the scheduling solution set, so as to put the processing plan represented by the optimal solution into the production process of the product.

[0025] Among them, the intelligent scheduling mechanism based on the hyperheuristic genetic algorithm combines a heuristic chromosome and a production order chromosome to form a scheduling solution. A heuristic chromosome consists of several heuristic operations, and a production order chromosome is the full permutation of the codes of all products to be processed. The intelligent scheduling mechanism based on the hyperheuristic genetic algorithm updates the scheduling solution by means of crossover and mutation operations between heuristic chromosomes, as well as genetic operations performed by the production order chromosome based on the heuristic operations in the heuristic chromosome.

[0026] A third aspect of 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 the intelligent scheduling method for production processes based on a hyperheuristic genetic algorithm as described above.

[0027] A fourth aspect of the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the intelligent scheduling method for production processes based on a hyperheuristic genetic algorithm as described above.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] This invention provides an intelligent scheduling method for production processes based on a hyperheuristic genetic algorithm. It adopts a method of encoding and optimizing low-level heuristic operation mechanisms in high-level heuristic chromosomes, and then using low-level heuristic operations to evolve scheduling solutions. This achieves self-learning evolution of the genetic operation mechanism, significantly improving the scheduling method's ability to evolve scheduling results and enhancing the robustness of scheduling performance. It not only exhibits more significant advantages in scheduling optimization performance, but also ensures that this method has broad applicability to different optimization objectives in different scheduling scenarios.

[0030] This invention provides an intelligent scheduling method for the production process based on a hyperheuristic genetic algorithm. It adopts a weighted summation method, which comprehensively considers the dominance degree of an individual in the population and the distribution density of neighborhood solutions. Based on the comprehensive dominance degree of individuals, it selects the dominant population. Compared with the method of selecting elite population based on non-dominant sorting strategy, it can comprehensively evaluate the degree of individual dominance and further maintain the characteristic diversity of the offspring population. Attached Figure Description

[0031] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0032] Figure 1 This is a flowchart of the intelligent scheduling method for the production process based on a hyperheuristic genetic algorithm according to Embodiment 1 of the present invention;

[0033] Figure 2 This is a diagram showing the relationship between high-level heuristics, low-level heuristics, and the scheduling problem domain in Embodiment 1 of the present invention.

[0034] Figure 3 This is a Gantt chart of an example of a single-product zero-wait intermittent production process according to Embodiment 1 of the present invention;

[0035] Figure 4 This is a Gantt chart of an example of a multi-batch, zero-wait, intermittent production process for products according to Embodiment 1 of the present invention. Detailed Implementation

[0036] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0037] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0038] Example 1

[0039] This embodiment provides an intelligent scheduling method for the production process based on a hyperheuristic genetic algorithm, such as... Figure 1 As shown, the specific steps include the following:

[0040] Step 1: Determine the quantity of products to be processed and the processing information for each product, and clarify the objective function on which the scheduling optimization of the zero-wait-interruption production process is based;

[0041] Step 2: Number all processed products, and initialize and optimize the scheduling solution set of the decision product processing plan through the intelligent scheduling mechanism based on the super-heuristic genetic (GAHH) algorithm. The optimization ends when the evolution termination condition is met.

[0042] Step 3: Select the approximate optimal solution that meets the actual processing needs from the solution set finally obtained in Step 2, and put the processing plan represented by the scheduling solution into the actual production and processing process.

[0043] The quantity of products to be processed in step 1 and the processing information for each product are as follows:

[0044] n products are put into a production workshop consisting of m machines for processing. Each product is processed in the same order from the first machine to the m-th machine before its processing task is completed. The n products form a product set J = {J1, J2, ..., Jm}. n A set of m devices is M = {M1, M2, ..., M}. m Product J i In device M k The processing steps performed above are O ik Its processing time is p ik In zero-wait intermittent production, each product starts from the second process, and each process has a wait time of 0 seconds. ik The start time is the time of its previous process O i(k-1) The completion time. For example... Figure 3 As shown, the processing flow of a product consisting of three steps is used as an example to illustrate the entire process of zero-wait intermittent production of the product.

[0045] The objective function underlying the zero-wait interval production process scheduling optimization in step 1 is described as follows:

[0046] The scheduling optimization objectives for a zero-wait-intermittent production process involve the maximum production completion time, the total processing time, and the maximum delay time. The objective functions for these three scheduling optimization objectives are expressed as follows:

[0047] (1) Maximum production completion time refers to the time required to complete the processing of all products. Using maximum production completion time as the optimization objective is to ensure that the scheduling solution minimizes the time spent on the production process. The objective function for maximum production completion time can be expressed as:

[0048] f1 = T = max{c i |i=1,2,...,n} (1)

[0049] Among them, c i This represents the duration from time zero until the processing of product i is completed.

[0050] (2) The total processing time refers to the sum of the production completion times of all products. The significance of this optimization objective is to ensure that the scheduling solution can shorten the maximum production completion time while minimizing the processing time of each product, thereby reducing total energy consumption and operating costs. The objective function for the total processing time can be expressed as:

[0051]

[0052] (3) The maximum delay time represents the delay duration of the product with the most severe delays in processing and delivery. By optimizing this objective, the scheduling solution should shorten the delay time for completing and delivering processed products, ensuring that all products are completed and delivered on time as much as possible. The objective function for the maximum delay time is:

[0053] f3 = MLT = max{0,(c i -d i )|i=1,2,...,n} (3)

[0054] Where, d i This indicates the latest delivery date for product i.

[0055] The optimization directions for scheduling include single-objective optimization and multi-objective optimization, among which:

[0056] (1) Single-objective optimization focuses on scheduling optimization for a single objective. The goal of this optimization is to find an approximate optimal solution through intelligent algorithms that makes the function value of the objective as close as possible to its optimal value. Taking the scheduling optimization for the maximum production completion time as an example, the optimization direction can be expressed as follows:

[0057] f = min f1 (4)

[0058] (2) Multi-objective optimization involves scheduling optimization for multiple or all objectives. Since the optimization objectives are not independent, optimizing one objective can affect the optimization results of others, and objectives often have mutual constraints. To achieve multi-objective collaborative optimization, it is necessary to calculate the objective function of each scheduling solution, determine its position in the multi-dimensional objective space, and measure the overall advantage of each individual to judge its merits. Through the evolutionary process, the overall advantage of individuals in the offspring population tends towards a relatively ideal point, thereby optimizing the scheduling solution. After the evolutionary scheduling is completed, the final generation's advantageous solution set is taken as the approximate optimal solution set. When optimizing the three objectives of maximum production completion time, total processing time, and maximum delay time simultaneously, the optimization direction can be expressed as:

[0059] f=(min f1,min f2,min f3) (5)

[0060] Taking the scheduling of the zero-wait batch production process of N-acetylglucosamine as an example:

[0061] Table 1. Product Processing Information Table

[0062] batch <![CDATA[J1]]> <![CDATA[J2]]> <![CDATA[J3]]> <![CDATA[J4]]> <![CDATA[J5]]> <![CDATA[J6]]> <![CDATA[J7]]> <![CDATA[J8]]> <![CDATA[J9]]> <![CDATA[J 10 ]]> Seed production 11 13 16 9 11 12 15 14 17 12 Fermentation production 57 46 54 48 53 49 41 52 54 48 acidification 2 4 2 4 3 2 3 4 3 1 Ceramic membrane 5 7 6 3 6 7 5 4 2 2 flocculants 1 3 1 4 3 4 2 3 1 2 Plate and frame 2 6 4 8 1 4 3 1 7 2 RO membrane 5 2 8 1 7 4 6 3 2 6 medium pressure membrane 5 7 3 7 5 6 4 8 3 5 MVR Concentrate 6 20 12 3 4 18 6 23 15 17 hydrolysis 10 14 9 11 10 12 13 13 6 8 Plate and frame, materials 12 20 4 23 14 8 21 23 15 17 Concentrated Crystallization 20 13 26 24 16 23 20 17 19 22 filtration 8 5 12 10 9 8 12 13 5 11 Double cone, sieve 4 7 6 3 5 6 4 7 8 6

[0063] Table 1 records the processing information (unit: hours) for 10 batches of products. The scheduling task is to optimize the production process of these 10 batches of products with zero waiting time.

[0064] Each batch of processed products is considered as a whole. These 10 products are put into a production workshop consisting of 14 production processes. Each product follows the process steps in Table 1, starting from the first process of seed production and continuing until the double cone and screening processes are completed to complete the processing task. The 10 products form a product set J = {J1, J2, ..., J...} 10 The production equipment set M = {M1, M2, ..., M} consists of 14 processes. 14 Product J i In device M k The processing steps performed above are O ik Its processing time is p ik For example, process O performed by J1 on equipment M1. 11 For seed production, the processing time p 11 The value is 11. In the zero-wait intermittent production process, each product starts from the second process, and each process has an O... ik The start time is the time of its previous process O i(k-1) The completion time. For example, if O 11Processing begins at time zero and takes 11 hours to complete. Immediately afterwards, O... 12 In the 11th hour, O 11 Processing begins immediately after the material is transferred to M2 after processing is completed.

[0065] If the scheduling optimization of the N-acetylglucosamine zero-wait intermittent production process is a single-objective optimization, the objective is to maximize production completion time. This example will use an intelligent scheduling mechanism based on the hyperheuristic genetic algorithm (GAHH) to arrange a reasonable production plan, ensuring that all 10 batches of products can be completed in approximately the shortest possible time.

[0066] If the scheduling optimization of the N-acetylglucosamine zero-wait intermittent production process is a multi-objective optimization, the multi-objective scheduling optimization of the N-acetylglucosamine zero-wait intermittent production process involves three objectives: maximum production completion time, total processing time, and maximum delay time. In this embodiment, the latest delivery date for each product is set to 1.5 times the total processing time of all 14 processes for that product.

[0067] The intelligent scheduling mechanism based on the GAHH algorithm in step 2 is a hyperheuristic genetic algorithm mechanism. This mechanism evolves the scheduling solution by directly evolving the high-level heuristic, and then by using low-level heuristic operations within the high-level heuristic. It optimizes the scheduling objective by calculating the overall dominance of individuals in the population. The specific optimization process includes:

[0068] Step (1) Set the evolution parameters, including the maximum number of iterations (Ite). max Population size scheduling size Advanced heuristic chromosome length L h Advanced heuristic crossover probability P hc Advanced heuristic mutation probability P hm The low-level heuristic for the probability P of chromosome production order operation l .

[0069] In this embodiment, if it is a single-objective optimization scheduling of the N-acetylglucosamine zero-wait intermittent production process, the maximum number of iterations is Ite. max The population size is 300, and the scheduling population is Pop. size The value is 50, and the high-level heuristic chromosome length L h The advanced heuristic crossover probability P is 6. hc The high-level heuristic mutation probability P is 0.8. hm The probability P of the low-level heuristic on the production order chromosome is 0.6. l The overall advantage score is 0.8, and the weight α for the overall advantage score is 0.6.

[0070] In this embodiment, if it is a multi-objective optimization scheduling of the N-acetylglucosamine zero-wait intermittent production process, the maximum number of iterations is Ite. max The population size is 500, and the scheduling population is Pop. size The value is 100, and the high-level heuristic chromosome length L is... h The advanced heuristic crossover probability P is 6. hc The high-level heuristic mutation probability P is 0.8. hm The probability P of the low-level heuristic on the production order chromosome is 0.6. l The overall advantage score is 0.8, and the weight α for the overall advantage score is 0.6.

[0071] Step (2) Based on the set size of the scheduling population, randomly generate Pop. size An initialization scheduling solution. For example... Figure 2 As shown, each scheduling solution consists of a heuristic chromosome and a production order chromosome. Each high-level heuristic chromosome consists of L... h The high-level heuristic chromosome consists of several low-level heuristic operations, and the operations listed in Table 2 represent all the heuristic operations that can be included in a high-level heuristic chromosome. Each high-level heuristic chromosome contains L... h Each low-level heuristic operation can only be selected from Table 2. Any operation can be repeated within the same high-level heuristic chromosome. The production order chromosome is the full permutation of the codes for all products to be processed. Pop is randomly generated from the initial solution set. size The full permutation of the product codes is generated and then combined with a high-level heuristic chromosome to form an initial scheduling solution.

[0072] Table 2. Heuristic Operation Table

[0073]

[0074]

[0075] In this embodiment, for a single-objective optimization scheduling of the N-acetylglucosamine zero-wait-interval production process, 50 initial scheduling solutions are randomly generated based on the set scheduling population size. Each scheduling solution consists of a heuristic chromosome and a production order chromosome. The six low-level heuristic operations in each high-level heuristic chromosome consist of operations LLH_1-LLH_6 from Table 2. Any operation can be repeated within the same high-level heuristic chromosome. For example, a valid high-level heuristic chromosome can be:

[0076] Chro high =[LLH_3,LLH_1,LLH_3,LLH_6,LLH_5,LLH_1] (6)

[0077] The production order chromosome is a complete permutation of the codes for all processed products. In the production order chromosome, each product's code can only appear once. For example, a production order chromosome that legally codes 10 products can be represented as:

[0078] Chro low =[5,9,10,6,1,8,3,7,2,4] (7)

[0079] In the initial solution set, 50 valid high-level heuristic chromosomes and 50 valid production order chromosomes are randomly generated, and one high-level heuristic chromosome and one production order chromosome are combined to form an initial scheduling solution.

[0080] In this embodiment, for the multi-objective optimization scheduling of the N-acetylglucosamine zero-wait-intermittent production process, 100 initial scheduling solutions are randomly generated according to the set scheduling population size. From the initial solution set, 100 valid high-level heuristic chromosomes and 100 valid production order chromosomes are randomly generated, and one high-level heuristic chromosome and one production order chromosome are combined to form an initial scheduling solution.

[0081] Step (3) Scheduling solutions perform genetic operations on the production order chromosomes based on the heuristic operation codes in the higher-level heuristic chromosomes. In each scheduling solution, the genetic operations are performed sequentially according to the order of the heuristic operation codes in the higher-level heuristic chromosomes. First, a random number in the interval [0,1] is generated. If this random number is less than or equal to P... l If the first operation in the higher-order heuristic chromosome is selected, the corresponding genetic operation is performed on the production order chromosome according to the operation method corresponding to the first operation encoded in Table 2; otherwise, the operation is not performed on the production order chromosome. Similarly, subsequent heuristic operations are performed sequentially in this manner. This continues until the L chromosome in the current higher-order heuristic chromosome is selected. h After all operations have been performed, the genetic operations on the production order chromosomes of the current scheduling solution are terminated. Pop in the population... size Each scheduling solution performs the genetic operations on the production order chromosome in the manner described above.

[0082] The higher-order heuristic chromosome Chro represented by equation (6) high For the production order chromosome represented by equation (7) low Taking the operation as an example. Chro high The first operation is LLH_3, which is a pre-processing operation: randomly select two codes from the production order chromosomes and place the latter code before the former. First, generate a random number in the interval [0,1]. If this random number is less than or equal to P... lThen, according to the corresponding operation method, the LLH_3 operation is performed on the production order chromosome. For example, if the randomly generated number is 0.5 and the randomly selected product codes are J2 and J6, then after performing the LLH_3 operation, the Chro shown in equation (7) will be... low The chromosome changes to:

[0083] Chro low = [5,9,10,2,6,1,8,3,7,4] (8)

[0084] However, if the randomly generated number is greater than 0.8, the current LLH_3 operation is not performed on the production order chromosome. Similarly, subsequent heuristic operations are performed sequentially in this manner. The genetic operations on the production order chromosome for the current scheduling solution are terminated after all six operations in the current high-level heuristic chromosome have been performed in the above manner. The genetic operations on the production order chromosome are performed on all other scheduling solutions in the population in the same manner.

[0085] Step (4) Higher-order chromosome crossing operation. Pop... size Each scheduling solution is randomly paired up to generate a Pop. size / 2 corresponding crossover combinations, and sequentially perform crossover operations on the high-level heuristic chromosome in each scheduling solution. First, generate a random number in the interval [0,1]. If this random number is less than or equal to P... hc If the condition is met, a segment is randomly selected from one of the higher-level heuristic chromosomes in the current crossover combination. The position and length of this segment are randomly determined, and it is then exchanged with a segment at the corresponding position in another higher-level heuristic chromosome in the combination. Otherwise, the crossover operation is not performed on the current combination. The same method is used for all other crossover combinations.

[0086] For example, there is a pair of higher-order heuristic chromosomes as shown in equation (9):

[0087]

[0088] If a crossover operation is performed, first generate a random number in the interval [0,1]. If this random number is less than or equal to P... hc Then, a segment is randomly selected from one of the higher-level heuristic chromosomes in the current crossover combination. The position and length of this segment are randomly determined, and it is exchanged with a segment at the corresponding position in another higher-level heuristic chromosome in the combination. If the generated random number is 0.3, and in Chro... high1 The selected crossover segment begins at the second element of the chromosome and has a length of 3. Therefore, Chro... high1 With Chro high2 The chromosomes after crossing over are shown in equation (10):

[0089]

[0090] If the random number is greater than 0.8, this operation is not performed on that crossover combination. The same procedure is followed for the crossover combinations of the remaining groups.

[0091] Step (5) Higher-order chromosome variation operations. For Pop in the population... size The high-level heuristic chromosomes of each scheduling solution are mutated sequentially. A random number in the interval [0,1] is generated. If this random number is less than or equal to P... hm If the mutation occurs, an operation code is randomly selected from the high-level heuristic chromosome of the current scheduling solution and replaced with any of the remaining operation codes in Table 1; otherwise, no mutation operation is performed on the current high-level heuristic chromosome. The mutation operation is performed in the same way on the high-level heuristic chromosomes of the other scheduling solutions.

[0092] For example, to modify Chro in equation (10) high1 'The mutation operation was performed, generating a random number 0.6. The randomly selected mutation code is the 5th element in the chromosome, and the randomly generated replacement code is LLH_2. The mutated chromosome is shown in equation (11):'

[0093] Chro high1 ”=[LLH_3,LLH_5,LLH_1,LLH_5,LLH_2,LLH_1] (11)

[0094] If the random number is greater than 0.6, do not perform a mutation operation on the current high-level heuristic chromosome. Perform the mutation operation using the same steps on all other high-level heuristic chromosomes for the remaining scheduling solutions.

[0095] Step (6) generates a combined parent and offspring population. The scheduling solutions that have undergone sequential production order chromosome genetic operations, higher-order chromosome crossover operations, and higher-order chromosome mutation operations are used as offspring individuals. The original individuals are used as their parent individuals. Pop size Parental individual and Pop size Each offspring individual forms a group of size A combined population.

[0096] Step (7) Calculation of objective function value. Based on the production order in each scheduling solution, and according to the constraints of the zero-wait-intermittent production process, the corresponding processing scheme is generated, and the corresponding objective function value is calculated.

[0097] In this embodiment, if it is a single-objective optimization scheduling of the N-acetylglucosamine zero-wait intermittent production process, the corresponding maximum production completion time is calculated according to the objective function f1.

[0098] In this embodiment, if it is a multi-objective optimization scheduling of the N-acetylglucosamine zero-wait intermittent production process, the results of the three objectives, namely the maximum production completion time, the total processing time, and the maximum delay time, are calculated according to the objective functions f1, f2, and f3.

[0099] Step (8) Select the dominant solution based on the individual comprehensive dominance measurement algorithm.

[0100] If it is a single-objective optimization scheduling, then the overall dominance of each individual in the joint population in that objective dimension is calculated using an individual comprehensive dominance measurement algorithm, and the top Pop with the highest dominance is selected. size Each scheduling solution is considered a dominant individual and forms a dominant solution set.

[0101] For the single-objective optimization scheduling of the N-acetylglucosamine zero-wait intermittent production process, the maximum production completion time of the 100 scheduling solutions in the parent-child joint population is sorted from smallest to largest. The number of individuals in the joint population whose objective value is better than each scheduling solution is calculated and converted into a normalized value. Then, the neighborhood solution distribution density of each individual in the objective dimension is calculated. Finally, the overall dominance of the individuals is calculated by weighting, as shown in Equation (12). Based on the results, the individuals in the population are re-sorted, and the top 50 individuals with the best overall dominance are selected.

[0102] wd x =α·gn x / (Pop size ·2)+(1-α)·cd x (12)

[0103] Among them, wd x gn represents the overall advantage of scheduling solution x. x cd is the number of solutions in the joint population whose objective value is better than the scheduling solution x. x Let be the neighborhood distribution density of individual x, and α be the weighting coefficient. x The calculation method is as follows:

[0104] cd x =lcd x +rcd x (13)

[0105]

[0106]

[0107] Among them, LCD x and RCD x These represent the left and right proximity of the scheduling solution x, respectively; ly x and ry xRepresent the distances between scheduling solution x and its left and right adjacent solutions, respectively; y max and represent the maximum distance between adjacent solutions in the joint population.

[0108] For multi-objective optimization scheduling, an individual comprehensive dominance algorithm is used to calculate and select the top 100 scheduling solutions with the highest comprehensive dominance in the multi-objective state space as dominant solutions, forming a dominant solution set. The individual comprehensive dominance of a scheduling solution is expressed as a weighted sum of the individual's dominance and the distribution density of its neighborhood solutions in the current joint population, calculated as follows:

[0109] wd x =α·nd x +(1-α)·cd x (16)

[0110] Among them, wd x nd represents the overall dominance of individual x. x cd represents the degree of dominance of individual x. x Let be the neighborhood distribution density of individual x, and α be the weighting coefficient. Dominance degree refers to the normalized result of an individual's non-dominance level in a joint population. The dominance degree is calculated as follows:

[0111]

[0112] Among them, np x np represents the non-dominated level of the scheduling solution x, i.e., the number of solutions that dominate x in the joint population. max The non-dominated level is the highest level of dominance for the solution in the joint population. The neighborhood solution distribution density represents the combined level of proximity between this individual and its neighbors in each target dimension within the distribution space of the joint population. The neighborhood solution distribution density is calculated as follows:

[0113]

[0114]

[0115]

[0116]

[0117] In equation (18), N is the number of targets; This indicates that the scheduling solution x is in the target f g The degree of proximity to neighboring individuals in a dimension. In equation (19), and These represent the scheduling solution x in the target f, respectively. g The degree of left and right proximity in the dimension. In equations (20) and (21), and These represent the scheduling solution x in the target f, respectively. g Distance to the left and right adjacent solutions in the dimension; and Let f represent the target population in the population. g The maximum and minimum values ​​are calculated. The smaller the overall dominance score, the higher the dominance ranking of the scheduling solution. Finally, based on the ranking results, the top 100 individuals are selected as dominant individuals, forming the dominant solution set for this generation.

[0118] Step (9) If the number of iterations has not yet reached the maximum number of iterations (Ite) max If the current dominant solution set is selected, return to step (3) and use the current dominant solution set as the parent population for the next round of genetic operations; otherwise, end the evolutionary process.

[0119] Step (10) outputs the production order chromosome in the last generation of advantageous solutions as the scheduling result.

[0120] For the multi-objective optimization scheduling of the N-acetylglucosamine zero-wait intermittent production process, the method (GAHH) in this invention is compared with the NSGA-II algorithm (derived from the paper "A fast and elitist multiobjective genetic algorithm: NSGA-II" published in IEEE Transactions on Evolutionary Computation in 2002) and the PSO algorithm (derived from the paper "A comprehensive survey on particle swarm optimization algorithm and its applications" published in Mathematical Problems in Engineering in 2015). By testing and comparing the performance of the non-dominated solution sets obtained by different algorithms in the same multi-objective scheduling scenario, the advantages and disadvantages of this invention in achieving zero-wait intermittent production process scheduling optimization are analyzed.

[0121] Here, the Pareto front solution set obtained by the method of this invention in this embodiment is compared and analyzed with the Pareto front solution sets obtained by two comparison methods using the coverage set measure, namely the C measure (derived from the paper Comparison of multiobjective evolutionary algorithms: empirical results published in Evolutionary Computation in 2000). After 10 independent repeated experiments, the average value of the C measure results is shown in Table 3.

[0122] Table 3. Results of C-measure

[0123]

[0124] Test results show that the scheduling solution set obtained by the intelligent scheduling mechanism based on the GAHH algorithm used in this invention has a more significant advantage than the solution set obtained by the comparison algorithm.

[0125] The method for selecting an approximate optimal solution in step 3 is described as follows:

[0126] When scheduling optimization becomes a single-objective optimization, the scheduling solution with the optimal objective function value is selected from the last generation of advantageous solutions as the approximate optimal solution. Based on the production order determined by this solution, and under the constraint of a zero-wait-intermittent production process, the corresponding processing plan is generated and put into actual production.

[0127] For the single-objective optimization scheduling of the N-acetylglucosamine zero-wait intermittent production process, the scheduling result should select the scheduling solution with the best objective function value in the last generation of advantageous solutions, that is, the scheme with the shortest maximum production completion time. In this example, the approximate optimal scheme for minimizing the maximum production completion time for 10 batches of N-acetylglucosamine zero-wait intermittent production process provided by the intelligent scheduling mechanism based on the hyperheuristic genetic (GAHH) algorithm is shown in Equation (22):

[0128] Chro low_best =[4,3,9,7,6,8,5,2,10,1] (22)

[0129] Based on this processing sequence, 10 batches of N-acetylglucosamine were produced with zero waiting time intermittently, resulting in a maximum production completion time of 591 hours. This scheme is significantly superior to the processing plan scheme manually designed by the scheduler as shown in equation (23):

[0130] Chro low_manual =[8,6,1,5,2,3,9,4,10,7] (23)

[0131] The maximum production completion time for this plan is 615 hours.

[0132] If scheduling optimization is a multi-objective optimization, then each optimization objective needs to be assigned a corresponding weight according to actual needs, and the weighted objective function value should be calculated. The scheduling solution with the best result is then used as the approximate optimal solution and implemented in the actual production process. The weighted objective is calculated according to the following formula:

[0133]

[0134]

[0135] Among them, f w For the weighted objective function, ω g For the g-th target f g The weight.

[0136] The weights of each objective can be subjectively assigned based on actual needs. This embodiment first obtains subjective weights through subjective decision analysis based on the Analytic Hierarchy Process (AHP), then derives objective weights through objective decision analysis based on the entropy weight method, and finally uses the Nash equilibrium model from game theory to calculate the comprehensive weights based on both subjective and objective weights.

[0137] ω=(ω1,ω2,ω3)=(0.329,0.328,0.343) (26)

[0138] Finally, the approximate optimal solution obtained by multi-objective scheduling optimization for the zero-wait intermittent production process of N-acetylglucosamine in this embodiment is shown in Equation (27):

[0139] Chro low_best =[4,7,2,10,6,8,5,3,9,1] (27)

[0140] Based on this processing sequence, 10 batches of N-acetylglucosamine were produced with zero waiting time intermittently. The maximum production completion time was 591, the total processing time was 3706, and the maximum delay time was 369.

[0141] This embodiment takes the N-acetylglucosamine production process as an example, and uses the GAHH algorithm to implement single-objective and multi-objective scheduling optimization for the production process, providing a scientific and efficient production scheme. The results show that the scheduling optimization process based on the GAHH algorithm for zero-wait intermittent production processes provided by this invention has stable performance, and its results are significant. As a hyperheuristic optimization mechanism, the GAHH algorithm enables the scheduling optimization process to no longer rely excessively on the manual design of scheduling experts, demonstrating the advantages of knowledge automation and making the method of this invention more widely applicable.

[0142] This application provides an efficient, stable, robust, and applicable optimization scheduling method for the zero-wait-intermittent production process scheduling problem. Considering the limitations of metaheuristic algorithm improvements, which often only address specific scheduling scenarios and optimization objectives, and their reliance on expert knowledge, this invention uses the zero-wait-intermittent production process as the scheduling scenario. It employs a method that encodes and optimizes low-level heuristic operation mechanisms in a high-level heuristic chromosome, and then uses these low-level heuristic operations to evolve the scheduling solution. This hyperheuristic evolutionary approach, through the concept of knowledge automation, achieves self-learning evolution of the genetic operation mechanism, significantly improving the scheduling method's ability to evolve scheduling results and enhancing the robustness of scheduling performance. This method not only exhibits significantly superior scheduling optimization performance but also ensures broad applicability to different optimization objectives in various scheduling scenarios.

[0143] Furthermore, this invention improves upon traditional methods for selecting elite populations based on non-dominated ranking strategies by employing a weighted summation approach. This approach comprehensively considers both an individual's dominance within the population and the density of neighborhood solutions, selecting the dominant population based on the individual's overall dominance. Compared to methods based on non-dominated ranking strategies, the selection method in this invention can comprehensively evaluate the degree of individual dominance and further maintain the characteristic diversity of the offspring population.

[0144] Example 2

[0145] This embodiment provides an intelligent scheduling system for production processes based on a hyperheuristic genetic algorithm, which specifically includes:

[0146] The processing plan optimization module is configured to: obtain the quantity of products to be processed and the processing information of each product, and initialize and optimize the scheduling solution set of the product processing plan through an intelligent scheduling mechanism based on a hyperheuristic genetic algorithm;

[0147] The optimal solution selection module is configured to select the optimal solution from the scheduling solution set, so as to put the processing plan represented by the optimal solution into the production process of the product.

[0148] Among them, the intelligent scheduling mechanism based on the hyperheuristic genetic algorithm combines a heuristic chromosome and a production order chromosome to form a scheduling solution. A heuristic chromosome consists of several heuristic operations, and a production order chromosome is the full permutation of the codes of all products to be processed. The intelligent scheduling mechanism based on the hyperheuristic genetic algorithm updates the scheduling solution by means of crossover and mutation operations between heuristic chromosomes, as well as genetic operations performed by the production order chromosome based on the heuristic operations in the heuristic chromosome.

[0149] It should be noted that each module in this embodiment corresponds one-to-one with each step in Embodiment 1, and their specific implementation processes are the same, so they will not be repeated here.

[0150] Example 3

[0151] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the intelligent scheduling method for the production process based on a hyperheuristic genetic algorithm as described in Embodiment 1 above.

[0152] Example 4

[0153] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the intelligent scheduling method for production processes based on a hyperheuristic genetic algorithm as described in Embodiment 1 above.

[0154] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0155] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0156] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0157] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0158] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0159] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A production process intelligent scheduling method based on a hyperheuristic genetic algorithm, characterized in that, include: The quantity of products to be processed and the processing information of each product are obtained. The scheduling solution set of the product processing plan is initialized and optimized through an intelligent scheduling mechanism based on a hyperheuristic genetic algorithm. The optimal solution is selected from the set of scheduling solutions, and the processing plan represented by the optimal solution is put into the production process of the product. Among them, the intelligent scheduling mechanism based on the hyperheuristic genetic algorithm combines a heuristic chromosome and a production order chromosome to form a scheduling solution. A heuristic chromosome consists of several heuristic operations, and a production order chromosome is the full permutation of the codes of all products to be processed. The intelligent scheduling mechanism based on the hyperheuristic genetic algorithm updates the scheduling solution by means of crossover and mutation operations between heuristic chromosomes, as well as genetic operations performed by the production order chromosome based on the heuristic operations in the heuristic chromosome. For multi-objective optimization scheduling, the optimization objectives of the intelligent scheduling mechanism based on hyperheuristic genetic algorithm are maximum production completion time, total processing time and maximum delay time; In each iteration, the intelligent scheduling mechanism based on the hyperheuristic genetic algorithm selects several scheduling solutions with the highest comprehensive advantage from the combined parent and offspring population as dominant individuals, forming a dominant solution set, and uses the selected dominant solution set as the parent population for the next round of genetic operations. For multi-objective optimization scheduling, the overall dominance is a weighted sum of the dominance and the neighborhood solution distribution density, and the formula for calculating the overall dominance is: in, Indicates the scheduling solution The overall advantages For the objective value in the joint population to be better than the scheduling solution The number of solutions, For individuals The neighborhood solution distribution density, These are the weighting coefficients.

2. The intelligent scheduling method for production processes based on a hyperheuristic genetic algorithm as described in claim 1, characterized in that, The heuristic operations include one or more of the following: random swap, adjacent swap, pre-operation, post-operation, reverse operation, and truncation operation.

3. The intelligent scheduling method for production processes based on a hyperheuristic genetic algorithm as described in claim 1, characterized in that, The processing information of the product includes: the processing steps and processing time performed on each device.

4. The intelligent scheduling method for production processes based on hyperheuristic genetic algorithms as described in claim 1, characterized in that, For multi-objective optimization scheduling, the optimal solution is selected from the scheduling solution set based on the weighted objective function value; The weighted objective function value is: in, For the first g One goal The weight.

5. A production process intelligent scheduling system based on a hyperheuristic genetic algorithm, characterized in that, include: The processing plan optimization module is configured to: obtain the quantity of products to be processed and the processing information of each product, and initialize and optimize the scheduling solution set of the product processing plan through an intelligent scheduling mechanism based on a hyperheuristic genetic algorithm; The optimal solution selection module is configured to select the optimal solution from the scheduling solution set, so as to put the processing plan represented by the optimal solution into the production process of the product. Among them, the intelligent scheduling mechanism based on the hyperheuristic genetic algorithm combines a heuristic chromosome and a production order chromosome to form a scheduling solution. A heuristic chromosome consists of several heuristic operations, and a production order chromosome is the full permutation of the codes of all products to be processed. The intelligent scheduling mechanism based on the hyperheuristic genetic algorithm updates the scheduling solution by means of crossover and mutation operations between heuristic chromosomes, as well as genetic operations performed by the production order chromosome based on the heuristic operations in the heuristic chromosome. For multi-objective optimization scheduling, the optimization objectives of the intelligent scheduling mechanism based on hyperheuristic genetic algorithm are maximum production completion time, total processing time and maximum delay time; In each iteration, the intelligent scheduling mechanism based on the hyperheuristic genetic algorithm selects several scheduling solutions with the highest comprehensive advantage from the combined parent and offspring population as dominant individuals, forming a dominant solution set, and uses the selected dominant solution set as the parent population for the next round of genetic operations. For multi-objective optimization scheduling, the overall dominance is a weighted sum of the dominance and the neighborhood solution distribution density, and the formula for calculating the overall dominance is: in, Indicates the scheduling solution The overall advantages For the objective value in the joint population to be better than the scheduling solution The number of solutions, For individuals The neighborhood solution distribution density, These are the weighting coefficients.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the intelligent scheduling method for production processes based on hyperheuristic genetic algorithms as described in any one of claims 1-4.

7. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the intelligent scheduling method for production processes based on hyperheuristic genetic algorithms as described in any one of claims 1-4.

Images