Decision tree-based uncertainty remanufacturing scheduling method

Abstract

The invention discloses an uncertainty remanufacturing scheduling method based on a decision-making tree, which aims to maximize the remanufacturing total profit and minimize the total time required for remanufacturing all waste products, sets the quality grade of the waste products as a fuzzy variable, evaluates the remanufacturability of the waste products by adopting the decision-making tree, and finally achieves the remanufacturing scheduling of the waste products. And a remanufacturing line with the maximum expected profit for remanufacturing the waste products is selected. Then, an improved teaching and learning optimization algorithm is adopted, learners adopt self-adaptive inertia weight to update and reprocess an operation sequence in a teacher stage, and when falling into local optimum, the learners learn to the teacher, the optimal learner and the learner with the maximum adaptive value improvement in the last iteration at the same time; local optimum of the algorithm is relieved; and the diversity of the population is improved. According to the method, efficient solving of the model is realized, and a more reasonable scheduling scheme can be obtained.

Description

Technical Field

[0001] The present application belongs to the technical field of remanufacturing scheduling, and in particular, relates to an uncertainty remanufacturing scheduling method based on a decision tree. Background Art

[0002] Remanufacturing is a recyclable manufacturing process based on the original manufacturing process, which aims to develop the residual value of waste products. With the continuous development of the remanufacturing industry, the remanufacturing scheduling problem has attracted more and more attention from academia and industry.

[0003] Due to the different usage levels of waste products and the efficiency of remanufacturing machines, there is significant uncertainty in the remanufacturing scheduling process, which has a significant impact on the evaluation of the remanufacturability of waste products and the selection of remanufacturing lines. Remanufacturability refers to the potential for recycled waste products to be remanufactured, which can be described as an assessment of whether the waste products are worth remanufacturing. Many factors may have a positive or negative impact on the remanufacturability of waste products, such as the profit generated by scrapping or remanufacturing waste products. After different levels of use, the quality of waste products varies significantly, affecting the profit generated by scrapping or remanufacturing waste products, and thus affecting the remanufacturability of waste products. At the same time, when remanufacturing waste products of different qualities, remanufacturing lines with different reprocessing operations can be selected. The time required to remanufacture waste products using different remanufacturing lines is different.

[0004] At present, most technical solutions for remanufacturing scheduling only consider deterministic factors, ignoring the impact of uncertain factors on remanufacturing scheduling. Some scholars have studied uncertain remanufacturing scheduling based on experience or random optimization methods, but have not clearly described uncertainty. Some scholars have used fuzzy methods such as fuzzy logic theory and linear fuzzy functions to transform uncertainty in remanufacturing scheduling into certainty, but have not explored the impact of uncertainty on evaluating the remanufacturability of waste products and selecting appropriate remanufacturing lines.

[0005] Since the impact of uncertainty factors on scheduling cannot be ignored, more and more scholars have begun to pay attention to uncertainty scheduling. Some scholars have proposed an uncertain remanufacturing production optimization scheduling system model constrained by two random variables. And using experience-based strategies, the reprocessing scheduling problem with uncertainty in reprocessing time in the remanufacturing system is solved; and using random simulation models to deal with the uncertainty of remanufacturing system scheduling. However, neither random optimization nor experience-based methods can clearly describe uncertainty, so they are only suitable for solving simple uncertainty scheduling problems.

[0006] Some scholars have also used fuzzy methods to solve the uncertain remanufacturing scheduling problem and used fuzzy logic theory to quantify the superposition effect of multiple uncertainties in the remanufacturing scheduling model. However, these methods only transform uncertainty into certainty through certain theories and methods to solve the uncertain remanufacturing scheduling problem, and have not explored the impact of uncertainty on remanufacturability evaluation and remanufacturing line selection. Summary of the invention

[0007] The purpose of this application is to provide a decision tree-based uncertainty remanufacturing scheduling method, which addresses the impact of uncertainty by constructing a decision tree and combining fuzzy numbers, and proposes remanufacturing strategies related to remanufacturability evaluation and remanufacturing line selection.

[0008] In order to achieve the above purpose, the technical solution of this application is as follows: An uncertain remanufacturing scheduling method based on a decision tree, comprising: Step 1: Create a remanufacturing optimization model with the goal of maximizing the total remanufacturing profit and minimizing the total time required to remanufacture all waste products; Step 2: Set the quality level of the waste products as a fuzzy variable, use a decision tree to evaluate the remanufacturability of the waste products, and select the remanufacturing line with the largest expected profit for remanufacturing the waste products as the remanufacturing line for the waste products; Step 3, using the remanufacturability evaluation results of the waste products and the selection results of the remanufacturing line as the first dimension of the scheduling plan, and the reprocessing operation sequence of the waste products on the remanufacturing line as the second dimension, to generate an initial population; Step 4: Determine whether it is trapped in a local optimum. If not, execute the teacher phase of the teaching and learning optimization algorithm. The learner uses the adaptive inertia weight to update the reprocessing operation sequence. If yes, execute the teacher phase of the teaching and learning optimization algorithm combined with the differential evolution algorithm. The learner learns from the teacher, the best learner, and the learner with the largest fitness improvement in the previous iteration at the same time, and updates the reprocessing operation sequence. Step 5: Execute the learner phase of the teaching and learning optimization algorithm, where the learner updates the reprocessing operation sequence; Step 6: Determine whether the termination condition is met. If not, return to step 4 to continue iterative optimization. After the termination condition is met, the optimal remanufacturing scheduling plan is obtained.

[0009] Furthermore, the first dimension includes three layers, the first layer represents the serial number of the waste products; the second layer represents the remanufacturability evaluation results of the waste products; the third layer represents the selection results of the remanufacturing line; The second dimension includes two layers, the first layer represents the real number coding of the reprocessing operation sequence of the waste product on the corresponding remanufacturing line, and the second layer represents the serial number of the reprocessing operation sequence of the waste product on the corresponding remanufacturing line.

[0010] Furthermore, the adaptive inertia weight is calculated by the following formula:

[0011] in, Represents the current iteration number, represents the maximum number of iterations, Representative The number of individuals whose fitness values ​​have improved in the iterations, Represents the total number of individuals in the population.

[0012] Further, the learner adopts adaptive inertia weight to update the reprocessing operation sequence, including: The real number code corresponding to the reprocessing operation sequence is calculated according to the following formula:

[0013] in, represents the i-th real number code corresponding to the reprocessing operation sequence of the new learner individual, represents the i-th real number code corresponding to the reprocessing operation sequence of the original learner individual, represents the i-th real number code corresponding to the reprocessing operation sequence of the teacher individual, represents the average value of the i-th real number code corresponding to the reprocessing operation sequence of all individuals in the group; represents the heuristic step length, which is a random number between 0 and 2; Compare the fitness values ​​of new learners with those of original learners, and retain learners with higher fitness values ​​to form a new population.

[0014] Furthermore, the learner simultaneously learns from the teacher, the best learner, and the learner with the largest fitness improvement in the previous iteration, and updates the reprocessing operation sequence, including: The learner updates the real number encoding corresponding to the reprocessing operation sequence according to the following formula:

[0015] in, represents the i-th real number code corresponding to the reprocessing operation sequence of the best learner, The i-th real number code corresponding to the reprocessing operation sequence of the learner with the largest fitness improvement in the previous iteration; Compare the fitness values ​​of new learners with those of original learners, and retain learners with higher fitness values ​​to form a new population.

[0016] Furthermore, the fitness value is calculated as follows: … …

[0017] in, Representative individual The fitness value of , Represents the total number of individuals in the population, TT n Represents an individual The corresponding total time required to remanufacture all waste products.

[0018] This application proposes an uncertain remanufacturing scheduling method based on a decision tree, which uses a decision tree to evaluate the remanufacturing ability of waste products and select remanufacturing lines. A new solution representation scheme is adopted, and an adaptive inertia weight adjustment function is used in combination with the differential mutation strategy of the differential evolution algorithm to alleviate the algorithm from falling into the local optimum and improve the diversity of the population, thereby achieving efficient solution of the model and obtaining a more reasonable scheduling scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of the uncertainty remanufacturing scheduling method based on decision tree for this application.

[0020] Figure 2 A solution of the remanufacturing optimization model of the present application represents an embodiment. DETAILED DESCRIPTION

[0021] 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.

[0022] Due to the different degrees of use of recycled waste products, there are significant differences in their quality levels. Remanufacturing waste products will consume certain production costs, and recycling failed remanufacturing products or selling successfully remanufactured products will obtain certain production income. The remaining part of the production income after deducting the production cost is the remanufacturing profit. The remanufacturing profit will vary with the quality level of the waste products. In addition, different remanufacturing lines generally contain different reprocessing operations. Therefore, the remanufacturing time and cost required to remanufacture the same waste product through different remanufacturing lines are also different.

[0023] The present application considers multiple non-equivalent parallel remanufacturing lines, namely, a high line, a medium line and a low line, which are respectively suitable for remanufacturing waste products of high, medium and low quality levels. The high line contains the simplest reprocessing operations, while the low line contains the most complex reprocessing operations. Although the high, medium and low remanufacturing lines can all be used to remanufacture waste products of the same quality level, the remanufacturability of the products and the time required for remanufacturing will vary depending on the selected remanufacturing line. Although the low line with the most complex reprocessing procedures can remanufacture high-quality waste products, the time required for remanufacturing is longer and the cost is higher, which in turn affects the remanufacturing profit and remanufacturability of the waste products.

[0024] The remanufacturing scheduling of this application is divided into two stages. The first stage uses a decision tree to evaluate the remanufacturability of waste products and select appropriate remanufacturing lines to maximize the total profit of remanufacturing. The second stage optimizes the reprocessing operation sequence of waste products on each remanufacturing line to minimize the total time required to remanufacture all waste products.

[0025] For the convenience of description, the symbols used in this application are described as follows: , the i-th waste product, ,in Indicates the total number of scrapped products.

[0026] , the jth quality level, , corresponding to high quality level, medium quality level and low quality level respectively.

[0027] , the zth remanufacturing line, , corresponding to the high line, middle line and low line of remanufacturing respectively.

[0028] , remanufacturing line The kth reprocessing unit of ,in Remanufacturing line Total number of remanufactured units.

[0029] , the total profit obtained from remanufacturing or scrapping all used products.

[0030] , the total time required to remanufacture all waste products.

[0031] , in the remanufacturing line Remanufactured products Profits obtained.

[0032] , in the remanufacturing line Remanufactured products The cost incurred.

[0033] , the remanufacturing quality level is Products The revenue generated.

[0034] , scrapped products Profits obtained.

[0035] , the scrap quality level is Products The revenue generated.

[0036] , reprocessing unit About Products The time when reprocessing starts.

[0037] , reprocessing unit About Products The time required for reprocessing.

[0038] , reprocessing unit About Products End time of reprocessing.

[0039] , a binary decision variable, 1 represents the product In the remanufacturing line Remanufacturing is performed on the , otherwise it is 0.

[0040] , a binary decision variable, 1 represents the product In the remanufacturing line The first one to be remanufactured, otherwise it is 0.

[0041] , a binary decision variable, 1 represents the product Worth remaking, otherwise 0.

[0042] , a binary decision variable, 1 represents the product The remanufacturing was successfully performed, otherwise it is 0.

[0043] An embodiment of the present application, such as Figure 1 As shown, a decision tree-based uncertainty remanufacturing scheduling method is provided, including: Step S1, creating a remanufacturing optimization model with the goal of maximizing the total remanufacturing profit and minimizing the total time required for remanufacturing all waste products.

[0044] In this embodiment, the total profit consists of two parts, namely, the profit obtained from remanufacturing waste products. and profits from scrapping of used products .

[0045] With the goal of maximizing the total profit of remanufacturing, the objective function created is shown in formula (1): (1) Profits from remanufacturing of used products and profits from scrapping of used products , which can be calculated by formula (2) and (3) respectively: (2) (3) In order to minimize the total time required to remanufacture all waste products, the objective function created is shown in formula (4): (4) Step S2: Set the quality grade of the waste products as a fuzzy variable, use a decision tree to evaluate the remanufacturability of the waste products, and select the remanufacturing line with the maximum expected profit for remanufacturing the waste products as the remanufacturing line for the waste products.

[0046] In this embodiment, the possible quality level of the waste product is set as a fuzzy variable , taking three quality levels as an example, it can be expressed by a fuzzy function, as shown in formula (5): (5) Among them, the probability Represents waste products Belong to the quality level The probability of , and satisfying formula (6): (6) Decision Trees for Evaluating Waste Products Remanufacturability, i.e. evaluating the remanufacturing of waste products Is it worth remanufacturing? As shown in formula (7): (7) in, Representatives in the remanufacturing line Products The expected profit from remanufacturing, Representatives of products The expected profit from scrapping can be calculated by formulas (8) and (9): (8) (9) in, and Representing the remanufacturing lines For quality level Waste products The success probability and failure probability of remanufacturing are calculated and satisfy formula (10): (10) if , that is, the product It is not worth remanufacturing and the product should be scrapped .if , that is, the product If it is worth remanufacturing, the decision tree is used again to select the appropriate remanufacturing line. By calculating and comparing the remanufactured products on different remanufacturing lines The expected profit can be obtained, so when the remanufacturing line About Products Expected profit from remanufacturing When the maximum value is reached, it is the most suitable remanufacturing line.

[0047] The decision tree consists of a decision diagram and possible results. This embodiment generates a tree structure based on the selection of different remanufacturing schemes, such as whether to remanufacture and on which remanufacturing line to remanufacture. The decision point represents the decision problem, the scheme branch represents the available schemes, and the probability branch represents the various possible results of the scheme. After calculating and comparing the expected profit values ​​under each branch (i.e., various remanufacturing schemes), a decision basis is provided and the optimal remanufacturing scheme is selected.

[0048] Step S3, generating an initial population with the remanufacturability evaluation results of the waste products and the selection results of the remanufacturing line as the first dimension of the scheduling plan and the reprocessing operation sequence of the waste products on the remanufacturing line as the second dimension.

[0049] The teaching and learning optimization algorithm is a group optimization algorithm proposed to solve engineering optimization problems. It mainly includes two stages, namely the teacher stage and the learner stage.

[0050] (1) Teacher stage: In this stage, the individual with the highest fitness value in the population is regarded as the teacher, and the remaining individuals are regarded as learners. Learners improve their own fitness value by continuously learning from the teacher, thereby improving the average fitness value of the entire population.

[0051] (2) Learner stage: In this stage, the learner improves its fitness value by interacting with another randomly selected learner.

[0052] The differential evolution algorithm is a heuristic random search algorithm, including initialization, mutation, crossover, selection and other stages. First, two individuals are randomly selected from the population as parent individuals for differential operation to obtain a differential vector (i.e., the difference in direction and magnitude between the two individuals in the solution space); then, the differential vector is superimposed on another individual to achieve mutation and generate a mutant individual; next, the two parent individuals and the mutant individuals are crossovered to generate new offspring individuals; finally, a selection operation is performed between the parent individuals and the offspring individuals, and the individuals that meet the requirements are retained as the next generation individuals of the population.

[0053] The teaching and learning optimization algorithm has the characteristics of strong solution space search capability, simple algorithm, few parameters, and strong adaptability, and is suitable for the optimization solution of complex problems. However, the basic teaching and learning optimization algorithm also has the disadvantages of poor population diversity and easy to fall into local optimality. The differential evolution algorithm can improve population diversity and help the teaching and learning optimization algorithm to jump out of the local optimality. Therefore, the present application proposes an adaptive hybrid optimization algorithm that combines the teaching and learning optimization algorithm with the differential evolution algorithm to efficiently solve the proposed remanufacturing optimization model (RSU-DT model). In the adaptive hybrid optimization algorithm, a new two-dimensional solution representation scheme is proposed to represent the solution of the RSU-DT model; an adaptive inertia weight adjustment function is used to help the algorithm jump out of the local optimality; and an improved teacher stage is proposed in combination with the differential evolution algorithm to improve the diversity of the population.

[0054] The remanufacturing scheduling problem proposed in this application is a complex discrete optimization problem, which consists of two parts: remanufacturing strategy selection and reprocessing operation sorting. Since the solution representation scheme of the basic teaching and learning optimization algorithm is one-dimensional and uses continuous real numbers to encode individuals, it cannot be directly applied to the optimization solution of the RSU-DT model. Therefore, this application proposes a new two-dimensional solution representation scheme to describe the individuals in the population. The RSU-DT model constructs a decision tree to propose a corresponding remanufacturing strategy. The first dimension corresponds to the remanufacturing evaluation of the waste products and the remanufacturing line selection results in the remanufacturing strategy. When the waste product is worth remanufacturing, the second dimension corresponds to the encoding of the waste product reprocessing operation sequence on each remanufacturing line.

[0055] In a specific embodiment, Figure 2 An example of a solution representation scheme is given.

[0056] like Figure 2As shown in part (a), the first dimension contains three layers: the first layer represents the serial number of the scrapped product; the second layer represents the result of the remanufacturing assessment of the scrapped product, that is, whether the scrapped product should be remanufactured; the third layer represents the selection result of the remanufacturing line, that is, which remanufacturing line is selected to remanufacture the scrapped product. Note that when the second layer of a scrapped product is "0 (scrapped)", its corresponding third layer is represented by 0, that is, the scrapped scrapped product does not need to select a remanufacturing line. For example, the first column represents the number of scrapped products on the remanufacturing line. Waste products The third column indicates the use of waste products To be scrapped.

[0057] like Figure 2 As shown in part (b), the second dimension represents the encoding of the reprocessing operation sequence of the scrapped product on the remanufacturing line. Note that the scrapped products that need to be scrapped do not need to be encoded in the second dimension. Since the teaching and learning optimization algorithm is only applicable to continuous real number encoding schemes, and the reprocessing operation sequence of scrapped products is discrete, this paper proposes an improved encoding scheme as follows.

[0058] The second dimension contains two layers. The first layer represents the real number encoding of the reprocessing operation sequence of the waste product on the corresponding remanufacturing line. When the initial population is generated, it is a continuous real number between 0 and 1 randomly generated. The second layer represents the number of the reprocessing operation sequence of the waste product on the corresponding remanufacturing line, which is a discrete value encoding. Taking the remanufacturing high line as an example, the smallest real number corresponds to the number of operations on the remanufacturing line. The scrap product with the smallest product number among the products to be reprocessed on the production line, and so on, until the remanufacturing line is completed. The reprocessing operation sequence is assigned to all remanufactured products on the remanufacturing line. The sequence of operations for reprocessing waste products is - - - ”.

[0059] Step S4, determine whether it is trapped in a local optimum. If not, execute the teacher stage of the teaching and learning optimization algorithm, and the learner uses adaptive inertia weights to update the reprocessing operation sequence. If so, execute the teacher stage of the teaching and learning optimization algorithm combined with the differential evolution algorithm. The learner learns from the teacher, the best learner, and the learner with the largest fitness improvement in the previous iteration at the same time, and updates the reprocessing operation sequence.

[0060] The basic teaching and learning optimization algorithm is prone to fall into the local optimum, while the inertia weight can alleviate the algorithm from falling into the local optimum by balancing the global and local solution space search capabilities of the algorithm. Therefore, this embodiment uses an adaptive inertia weight adjustment function to prevent falling into the local optimum.

[0061] The inertia weight of the current iteration It can be calculated by formula (11): (11) in, Represents the current iteration number, represents the maximum number of iterations, Representative The number of individuals whose fitness values ​​have improved in the iterations, Represents the total number of individuals in the population.

[0062] In order to alleviate the algorithm from falling into the local optimum and improve the diversity of the population, this embodiment proposes an improved teacher stage. A judgment of whether it is trapped in the local optimum is added. If it is not trapped in the local optimum, the learner uses the adaptive inertia weight to update the reprocessing operation sequence to prevent it from falling into the local optimum; if it has fallen into the local optimum, the teacher stage of the teaching and learning optimization algorithm combined with the differential evolution algorithm is executed. The learner learns from the teacher, the best learner, and the learner with the largest fitness value improvement in the previous iteration at the same time, and updates the reprocessing operation sequence to jump out of the local optimum.

[0063] To judge whether it is trapped in the local optimum, the fitness value of the individual after the previous 5 iterations can be calculated each time when judging. If the maximum fitness value of the current learner individual has not been improved, it is considered to be trapped in the local optimum.

[0064] When not falling into the local optimum, the learner uses adaptive inertia weights to update the reprocessing operation sequence, that is, to update the real number encoding corresponding to the reprocessing operation sequence according to formula (12): (12) in, represents the i-th real number code corresponding to the reprocessing operation sequence of the new learner individual, represents the i-th real number code corresponding to the reprocessing operation sequence of the original learner individual, represents the i-th real number code corresponding to the reprocessing operation sequence of the teacher individual, represents the average value of the i-th real number code corresponding to the reprocessing operation sequence of all individuals in the group; Represents the heuristic step size, which is a random number between 0 and 2.

[0065] When falling into a local optimum, the learner needs to learn from the teacher (i.e., the best individual in the population), the best learner (i.e., the suboptimal individual in the population), and the learner with the largest fitness improvement in the previous iteration. The learner updates the real number encoding corresponding to the reprocessing operation sequence according to formula (13): (13) in, represents the i-th real number code corresponding to the reprocessing operation sequence of the best learner, Represents the i-th real number encoding corresponding to the reprocessing operation sequence of the learner with the largest improvement in fitness in the previous iteration.

[0066] In this embodiment, the update reprocessing operation sequence will compare the fitness values ​​of the new learner individuals with the original learner individuals, and retain the learners with higher fitness values ​​to form a new population, where the individual The fitness value can be calculated according to formula (14): … … (14) in, Representative individual The fitness value of , Represents the total number of individuals in the population. n Represents an individual The corresponding total time required to remanufacture all waste products.

[0067] The fitness value of this embodiment reflects the difference between the maximum value of the total time required for all individuals to remanufacture all waste products and the total time required for this individual to remanufacture all waste products. The larger the fitness value, the shorter the total time required for this individual to remanufacture all waste products, so as to achieve the optimization goal of this application, that is, to minimize the total time required to remanufacture all waste products.

[0068] Step S5: Execute the learner phase of the teaching and learning optimization algorithm, and the learner updates the reprocessing operation sequence.

[0069] In the learner stage, the current learner randomly selects an individual from the population and learns from this individual to improve its own fitness. The learner updates the real number encoding corresponding to the reprocessing operation sequence according to formula (15):

[0070] (14) in, represents the i-th real number code corresponding to the reprocessing operation sequence of a randomly selected individual in the population, and Represent the fitness values ​​of the current individual and the randomly selected individual respectively.

[0071] Similarly, when updating the real number code corresponding to the reprocessing operation sequence, the fitness values ​​of the new learner individuals are compared with the original learner individuals, and the learners with higher fitness values ​​are retained to form a new population.

[0072] Step S6, determine whether the termination condition is met, if not, return to step S4 to continue iterative optimization, and after the termination condition is met, obtain the optimal remanufacturing scheduling plan.

[0073] The termination condition of the iteration can be to determine whether the preset number of iterations has been reached, or the maximum fitness value of the learner individual no longer increases, etc. If the termination condition is not met, the iteration is returned to continue. If the termination condition is met, it is considered that the optimal solution has been found, that is, the optimal remanufacturing scheduling plan has been found, and remanufacturing is subsequently arranged according to the scheduling plan.

[0074] In order to verify the above technical solution, the applicant conducted simulation experiments and comparative experiments on this method and other baseline algorithms to determine the parameters of the algorithm and evaluate the performance of the algorithm in solving the remanufacturing optimization model (RSU-DT model) of this application.

[0075] The baseline algorithms used in the experiment are differential evolution (DE), teaching and learning optimization (TLBO), discrete particle swarm optimization (DPSO) and extended particle swarm optimization (EPSO). The parameters in the simulation data set are randomly generated within a given interval. To ensure the reliability of the experiment, all experiments are run 5 times, and the total time required for remanufacturing of all waste products is used as the evaluation indicator.

[0076] The experiment was conducted on 14 simulation data sets to compare the proposed method with four other baseline algorithms. The experimental results are shown in Tables 1 and 2.

[0077] Table 1

[0078] Table 2

[0079] The experimental results in Table 1 show the solution space search performance of each algorithm when solving the model, with the average value and optimal value of TT as evaluation indicators, represented by "average value" and "optimal value" respectively. The experimental results in Table 2 show the stability and efficiency of the algorithm when solving the model, with the standard deviation of TT and the CPU running time of the algorithm as evaluation indicators, represented by "standard deviation" and "running time" respectively.

[0080] As shown in Table 1, on most data sets, the average and optimal values ​​of TT obtained by this method are better than those of other baseline algorithms, indicating that the solution space search performance of this method in solving the model is better than that of other baseline algorithms. As shown in Table 2, on most data sets, the standard deviation of TT obtained by this method is better than that of other baseline algorithms, indicating that this method has better stability in solving the model.

[0081] Remanufacturing has good economic value and social benefits, and remanufacturing scheduling is one of the key processes of remanufacturing. The actual remanufacturing scheduling process is always affected by a variety of uncertain factors, such as the uncertainty of the quality of scrap products and the uncertainty of the time required for product remanufacturing, which in turn reduces the economic value and social benefits of remanufacturing. This application proposes a new remanufacturing scheduling model, which uses a decision tree to realize the remanufacturability evaluation and remanufacturing line selection of scrap products. On this basis, a new solution algorithm is proposed, a new solution representation scheme is adopted, and an adaptive inertia weight adjustment function is used in combination with the differential mutation strategy of the differential evolution algorithm to alleviate the algorithm from falling into local optimality and improve the diversity of the population, thereby achieving efficient solution of the model.

[0082] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A decision tree-based uncertainty remanufacturing scheduling method, characterized in that: The uncertainty remanufacturing scheduling method based on a decision tree comprises: Step 1: Create a remanufacturing optimization model with the goal of maximizing the total remanufacturing profit and minimizing the total time required to remanufacture all waste products; Step 2: Set the quality level of the waste products as a fuzzy variable, use a decision tree to evaluate the remanufacturability of the waste products, and select the remanufacturing line with the largest expected profit for remanufacturing the waste products as the remanufacturing line for the waste products; Step 3, using the remanufacturability evaluation results of the waste products and the selection results of the remanufacturing line as the first dimension of the scheduling plan, and the reprocessing operation sequence of the waste products on the remanufacturing line as the second dimension, to generate an initial population; Step 4: Determine whether it is trapped in a local optimum. If not, execute the teacher phase of the teaching and learning optimization algorithm. The learner uses the adaptive inertia weight to update the reprocessing operation sequence. If yes, execute the teacher phase of the teaching and learning optimization algorithm combined with the differential evolution algorithm. The learner learns from the teacher, the best learner, and the learner with the largest fitness improvement in the previous iteration at the same time, and updates the reprocessing operation sequence. Step 5: Execute the learner phase of the teaching and learning optimization algorithm, where the learner updates the reprocessing operation sequence; Step 6: Determine whether the termination condition is met. If not, return to step 4 to continue iterative optimization. After the termination condition is met, the optimal remanufacturing scheduling plan is obtained.

2. The uncertainty remanufacturing scheduling method based on decision tree according to claim 1 is characterized in that: The first dimension includes three layers, the first layer represents the serial number of the waste products; the second layer represents the remanufacturability evaluation results of the waste products; the third layer represents the selection results of the remanufacturing line; The second dimension includes two layers, the first layer represents the real number coding of the reprocessing operation sequence of the waste product on the corresponding remanufacturing line, and the second layer represents the serial number of the reprocessing operation sequence of the waste product on the corresponding remanufacturing line.

3. The uncertainty remanufacturing scheduling method based on decision tree according to claim 1 is characterized in that: The adaptive inertia weight is calculated by the following formula: in, Represents the current iteration number, represents the maximum number of iterations, Representative The number of individuals whose fitness values ​​have improved in the iterations, Represents the total number of individuals in the population.

4. The uncertainty remanufacturing scheduling method based on decision tree according to claim 3 is characterized in that: The learner adopts adaptive inertia weights to update the reprocessing operation sequence, including: The real number code corresponding to the reprocessing operation sequence is calculated according to the following formula: in, represents the i-th real number code corresponding to the reprocessing operation sequence of the new learner individual, represents the i-th real number code corresponding to the reprocessing operation sequence of the original learner individual, represents the i-th real number code corresponding to the reprocessing operation sequence of the teacher individual, represents the average value of the i-th real number code corresponding to the reprocessing operation sequence of all individuals in the group; represents the heuristic step length, which is a random number between 0 and 2; Compare the fitness values ​​of new learners with those of original learners, and retain learners with higher fitness values ​​to form a new population.

5. The uncertainty remanufacturing scheduling method based on decision tree according to claim 1 is characterized in that: The learner simultaneously learns from the teacher, the best learner and the learner with the largest fitness improvement in the previous iteration, and updates the reprocessing operation sequence, including: The learner updates the real number encoding corresponding to the reprocessing operation sequence according to the following formula: in, represents the i-th real number code corresponding to the reprocessing operation sequence of the best learner, The i-th real number code corresponding to the reprocessing operation sequence of the learner with the largest fitness improvement in the previous iteration; Compare the fitness values ​​of new learners with those of original learners, and retain learners with higher fitness values ​​to form a new population.

6. The uncertainty remanufacturing scheduling method based on decision tree according to claim 4 or 5, characterized in that: The fitness value is calculated as follows: … … in, Representing an individual The fitness value of , Represents the total number of individuals in the population, TT n Represents an individual The corresponding total time required to remanufacture all waste products.

Images