Method for Setting up a Bilateral Local Disassembly Line Considering Electrical Constraints and Time-of-Use Electricity Prices

Abstract

The present invention provides a method for setting a bilateral local damage disassembly line considering electrical restrictions and time-of-use electricity prices, comprising the following steps: collecting the basic information of disassembly tasks and the unit electricity cost under different seasons and different time periods; establishing a bilateral local damage disassembly mixed integer programming model with the minimum number of workstations, the idle balance index, the disassembly profit and the disassembly energy consumption under different electricity consumption scenarios as the optimization objectives; solving the above-mentioned mixed integer programming model; meanwhile, the present invention also provides a new algorithm to improve the model solving efficiency. The present invention provides a new layout form and disassembly mode to solve problems such as non-disassemblable components, difficult product movement, and low efficiency encountered in the actual disassembly process of large and complex retired products, and at the same time introduces time-of-use electricity prices to find a balance between industrial efficiency and energy consumption, providing a basis for promoting efficient energy utilization.

Description

Technical Field

[0001] The present invention relates to the technical field of disassembly line balancing, and specifically to a bilateral partial disruption disassembly line setting method considering electrical restrictions and time-of-use electricity prices. Background Art

[0002] Due to the frequent occurrence of extreme weather events such as droughts, heavy rains, and high temperatures, the world is currently facing a severe energy crisis. The automotive manufacturing industry is a strategic pillar of the economy in many countries and has made a great contribution to the stable growth of the industry. However, as one of the industries with the largest consumption of metal resources, it poses a considerable challenge to the environment and energy consumption. Remanufacturing aims to use advanced technologies to restore or even improve the quality of end-of-life products, significantly reduce resource waste, alleviate the energy crisis, and promote the development of a circular economy. Disassembly is the prerequisite for remanufacturing. Efficient disassembly of end-of-life products can minimize environmental pollution by harmful substances and effectively recover components for subsequent remanufacturing, thus saving resources. To reduce costs and improve efficiency, many companies have established disassembly production lines, especially for large and complex retired automotive products. Effective allocation of disassembly tasks and load balancing, that is, the disassembly line balancing (DLB) problem, is crucial for improving disassembly efficiency. By solving this problem, enterprises can not only significantly improve the remanufacturing disassembly efficiency but also promote the development of a circular green economy.

[0003] Due to practical factors such as product size, failure mode, disassembly cost, and energy consumption, the disassembly of retired products is inherently complex. These factors pose continuous challenges to the research on the DLB problem. Unfortunately, current research mainly focuses on non-destructive disassembly, ignoring the influence of these factors, resulting in low efficiency and ultimately possible task failure. In addition, the disassembly industry is usually energy-intensive and tends to consume a large amount of energy. Therefore, most literature takes into account energy consumption and cost. However, these studies usually assume that the disassembly cost and energy consumption of each task are fixed and only focus on minimizing energy consumption. They rarely consider electricity costs and restrictions, especially under time-of-use electricity prices. This cannot be ignored in practice. Unfortunately, in the existing DLB problems, there is no work considering time-of-use electricity prices. Summary of the Invention

[0004] To solve the above problems, the main objective of the present invention is to provide a bilateral partial disruption disassembly line setting method considering electrical restrictions and time-of-use electricity prices, and also provide a solution method based on the improved water cycle algorithm (IWCA) to obtain a comprehensive and relatively optimal solution in multiple aspects, providing guidance and basis for the optimal setting of disassembly lines for large and complex retired products.

[0005] The choice of disassembly mode is usually related to the attributes of the end-of-life product itself. For example, Figure 1Task 41 among them is for the circuit board of scrapped LCD TVs. Due to its relatively high value, non-destructive disassembly is usually adopted to ensure its integrity. On the contrary, the recycling value of Task 5 is relatively low, and it can be quickly recycled through destructive disassembly. In addition, there are some components that cannot be disassembled non-destructively due to their own non-disassemblability. These components are often caused by some faults, such as aging, deformation, corrosion, etc. Therefore, only destructive disassembly can be carried out on these components. Therefore, introducing destructive disassembly is more suitable for practical engineering applications. In addition, the type of station layout of the disassembly line is also crucial for actual production. As Figure 1 shown in the disassembly line layout, this disassembly line can carry out disassembly work on both sides of the production line, so it is called a bilateral disassembly line. This layout can greatly reduce the handling length of retired products and improve the disassembly efficiency.

[0006] As is well known, as an energy-intensive industry, disassembly work consumes a large amount of electricity. However, most studies assume that the disassembly cost and energy consumption of each task are fixed, only focusing on minimizing energy consumption and ignoring cost savings under time-of-use electricity prices. As Figure 1 shown, the electricity consumption scenarios can be divided into off-peak, mid-peak, and peak periods. Time-of-use electricity prices result in higher electricity prices during peak periods and lower electricity prices during off-peak periods. This consideration is crucial in practice. Unfortunately, there is currently no study on DLB under time-of-use electricity prices.

[0007] The assumptions of the present invention are as follows: (1) Retired products are unique in type and there are a sufficient number of complete components, ignoring unplanned interruptions and task failures; (2) It is assumed that the operation proficiency of each worker is at the same level; (3) Each workstation is equipped with auxiliary lifting equipment to break the disassembly direction limit of parts; (4) All parts are allowed to be disassembled destructively and are equipped with disassembly tools; (5) The disassembly tool replacement time is included in the disassembly time of parts.

[0008] The technical solution of the present invention is:

[0009] A method for setting a bilateral partial destructive disassembly line considering electrical restrictions and time-of-use electricity prices, comprising the following steps:

[0010] S1. Obtain the basic information of the disassembly tasks, including the disassembly time under different disassembly modes, the recycling value of tasks under different disassembly modes, the unit power consumption of tasks under different disassembly modes, the tools required for disassembly, the destructive disassembly feasibility value, the priority relationship, and the unit electricity cost under different seasons and different time periods;

[0011] S2. Establish the multi-objective functions for the bilateral local destructive disassembly line, including the number of workstations, the balance rate of idle time, the disassembly profit considering time-of-use electricity price, and the disassembly energy consumption considering different disassembly modes, and establish the corresponding mixed-integer programming model; at the same time, establish the constraint conditions of the model, including the disassembly direction, priority, disassembly mode, time, location, and workstation of the tasks

[0012] S3. Solve the above-mentioned mixed-integer programming model;

[0013] In addition, the present invention proposes a new method for solving the above-mentioned mixed-integer programming model, which includes the following steps:

[0014] S31. Based on the double-layer coding method, perform two-stage decoding on it, and then generate the initial solution according to the Pareto elite strategy;

[0015] S32. Optimize and update the initial solution through the improved water cycle algorithm, and introduce the Pareto elite retention strategy to obtain the optimal solution and the optimal plan.

[0016] Among them, optimizing and updating the initial solution through the improved water cycle algorithm includes: after decoding is completed, generating the initial population, calculating the objective function, dividing the ocean, rivers, and streams, and generating the first-generation population; then performing ocean-river and river-stream position updates; then performing rainfall operations; combining the offspring population and the parent population to form a new generation of population and repeating the above operations until the number of iterations meets the termination conditions, and screening out the optimal solution through the hypervolume value (HV) as the final output.

[0017] Beneficial effects:

[0018] The present invention designs a method for setting a bilateral local destructive disassembly line considering electrical restrictions and time-of-use electricity price. Considering that the current research mainly focuses on non-destructive straight-line disassembly and ignores the situation where tasks cannot be disassembled due to the quality uncertainty of retired products, destructive disassembly is introduced to solve this problem. In addition, due to the difficulty of moving large and complex products, a bilateral layout is designed to reduce product movement by disassembling on both sides of the production line. In order to more accurately reflect the real-world situation, this study incorporates time-of-use electricity price into the model, which has not been considered in previous work. By exploring the impact of time-of-use electricity price on energy consumption and profit, a mixed-integer programming model is established, and the correctness of the model is verified through small-scale examples. In addition, the present invention proposes an improved water cycle algorithm (IWCA) to improve the model solving efficiency. The algorithm proposed by the present invention designs double-layer coding, two-stage decoding, two position update methods, and mutation operations based on the problem characteristics to efficiently solve the above-mentioned mixed-integer programming model. Description of the Drawings

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for the embodiments will be briefly introduced below.

[0020] Figure 1 is a bilateral partial disassembly line considering electrical restrictions and time-of-use electricity prices;

[0021] Figure 2 is a double-layer coding schematic diagram;

[0022] Figure 3 is a two-point crossover operation schematic diagram;

[0023] Figure 4 is a three-point crossover operation schematic diagram;

[0024] Figure 5 is a mutation operation schematic diagram;

[0025] Figure 6 is a three-dimensional view of the product engine;

[0026] Figure 7 is a box plot of multi-objective index evaluation when different algorithms are used to solve the automotive engine;

[0027] Figure 8 are 4 disassembly schemes for bilateral partial disassembly considering electrical restrictions and time-of-use electricity prices. Specific Embodiments

[0028] The present invention will be further described in detail below in conjunction with the embodiments and the accompanying drawings.

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0030] Explanation of symbols and decision variables:

[0031] Table 1 Explanation Table of Symbols and Decision Variables

[0032]

[0033]

[0034] To facilitate the understanding of the meanings of the embodiments and other formulas in the subsequent parts of the application document, Table 1 showing the symbols and decision variables is hereby given. Unless otherwise specified, the meanings of the same symbols and decision variables described hereinafter are the same as those in Table 1.

[0035] The present invention will be further described below in conjunction with embodiments:

[0036] In the following embodiments, unless otherwise specified, the operations described are conventional operations in the art.

[0037] A method for setting up a bilateral partial destructive disassembly line considering electrical restrictions and time-of-use electricity prices, the method comprising the following steps:

[0038] S1. Obtain the basic information of the disassembly tasks, including the disassembly time under different disassembly modes, the recovery value of the tasks under different disassembly modes, the unit power consumption of the tasks under different disassembly modes, the tools required for disassembly, the feasibility value of destructive disassembly, the priority relationship, and the unit electricity cost in different seasons and different time periods;

[0039] The present invention takes into account the impact of time-of-use electricity prices on the number of workstations, disassembly costs, and disassembly energy consumption. Each task can be selected from two disassembly modes. Due to different disassembly modes, the disassembly time, recovery value, unit power consumption, and unit electricity cost of each task are different.

[0040] S2. Establish a multi-objective function for the bilateral partial destructive disassembly line, including the number of workstations, the balance rate of idle time, the disassembly profit considering time-of-use electricity prices, and the disassembly energy consumption considering different disassembly modes, and establish a corresponding mixed-integer programming model; at the same time, establish the constraint conditions of the model, including constraints on the disassembly direction, priority, disassembly mode, time, location, and workstations of the tasks

[0041] The objective function of the constructed model is as follows:

[0042] Minimize the number of workstations:

[0043] Minimize the idle time balance:

[0044] Disassembly profit index considering time-of-use electricity prices:

[0045]

[0046] Among them, maximizing the disassembly profit requires considering multiple factors, including the disassembly income under different disassembly modes, the electricity cost of performing disassembly tasks under different modes, the electricity cost of turning on workstations, and the additional electricity cost of handling dangerous tasks.

[0047] Disassembly energy consumption indicators under different disassembly modes:

[0048]

[0049] Among them, the disassembly energy consumption indicators under different disassembly modes involve the power consumption of disassembly tasks, the additional energy consumption for handling dangerous tasks at the workstation, and the energy consumption of the workstation operation.

[0050] In the formula, S k is the workstation startup variable. If the k-th workstation is started, then S k = 1; otherwise, S k = 0; k is the workstation number, and W is the set of workstation numbers. CT is the cycle time; i represents the disassembly task number; N is the set of disassembly task numbers; l represents the left / right side index of the workstation, where l = 1 represents the left side and l = 2 represents the right side; L is the set of left / right side numbers of the workstation; t i represents the disassembly time of task i; G i is the disassembly revenue of the task; E i is the unit electricity cost; E k is the fixed unit time cost of the workstation; is the additional electricity cost generated by handling dangerous task i; h i is the task hazard attribute variable. If task i is a dangerous task, then h i = 1; otherwise, h i = 0; P i is the power consumption per unit disassembly time; P f is the fixed unit time power consumption of the workstation; P h is the power consumption per unit time for handling additional tasks.

[0051] Meanwhile, the relevant constraints of the model are expressed as follows:

[0052] Allocation constraints of disassembly tasks:

[0053]

[0054] The above formula indicates that each task is assigned to a workstation

[0055]

[0056] The above formula indicates that the task needs to satisfy the disassembly direction constraint. In the formula, D 1 represents the set of left-side tasks, and D 2 represents the set of right-side tasks.

[0057] Disassembly mode constraints:

[0058]

[0059] The above formula indicates that each task can only select one disassembly mode. Here, r represents the disassembly mode, r = 1 represents non-destructive disassembly, and r = 2 represents destructive disassembly. is the task disassembly mode variable, with the subscript i being the task number and the superscript r being the disassembly mode number. If task i adopts r disassembly modes, then Otherwise

[0060] {i ∈ DPe i > e max}

[0061]

[0062] The above two formulas indicate that only when the destructive disassembly feasibility value of a task is greater than the set threshold can the destructive disassembly mode be selected. In the formula, e i represents the destructive disassembly feasibility value of task i, which is known; e max is the set threshold, generally 0.5; DP is the set of destructive tasks.

[0063]

[0064] The above formula indicates that for tasks not belonging to the set of destructive tasks, both disassembly modes can be selected, but only one disassembly mode can be chosen.

[0065] {i ∈ NPh i + d i ≠ 0}

[0066]

[0067] The above formula indicates that dangerous tasks or tasks with requirements can only select non-destructive disassembly. In the formula, h i is the task hazard attribute variable. If task i is a dangerous task, then h i = 1, otherwise h i = 0; d i is the task requirement attribute variable. If task i is a task with requirements, then d i = 1, otherwise d i = 0; NP is the set of non-destructive tasks.

[0068] The relationship between disassembly time, recovery value, electricity cost, and disassembly mode:

[0069]

[0070] This constraint establishes a relationship among the disassembly time, benefit, and electricity cost of each task according to the disassembly mode used.

[0071] Workstation constraint:

[0072]

[0073] The above formula constrains the upper and lower bounds of the number of workstations. Where n is the total number of tasks

[0074]

[0075] The above formula means that when tasks are assigned to workstations, the workstations must be in the open state.

[0076] S k ≤S k-1 , k ∈ {2, ..., m}

[0077] The above constrains that the workstations can only be turned on sequentially

[0078] Location-related constraints:

[0079]

[0080] The above formula indicates that if tasks i and j are assigned to the same work position, they must occupy the same side of that work position.

[0081]

[0082] The above formula indicates that if tasks i and j are assigned to the same side of a work position, they cannot be located at different work positions:

[0083] Time-related constraints:

[0084]

[0085] The above formula means that the start times of all tasks must be greater than 0, ensuring that no task starts to execute before the production line starts running. Where w i represents the start time of task i

[0086]

[0087] The above formula means that any task assigned after another task must wait for the previous task to complete before starting disassembly.

[0088]

[0089] The above formula indicates that if two tasks i and j are assigned to the same side of the same workstation and task j is disassembled before task i, then task i must wait for task j to complete before starting to execute.

[0090]

[0091] The above formula indicates that all tasks must complete the disassembly work within the takt time assigned to their work positions.

[0092]

[0093] The above formula indicates that task i must wait for the subsequent task j to complete disassembly before it can start, even if they are on different sides of the same work station. In the formula, A ij = [a ij n×n , if task i is the immediate predecessor task of task j, then a ij = 1, otherwise a ij = 0.

[0094] S3. Based on the double-layer coding method, perform two-stage decoding on it, and then generate the initial solution according to the Pareto elite strategy; optimize and update the initial solution through the improved water cycle algorithm, including: after the decoding is completed, generate the initial population, calculate the objective function, divide the ocean, rivers, and streams to generate the first-generation population; then perform the ocean-river and river-stream position updates; then perform the rainfall operation; combine the offspring population and the parent population to form a new generation of population and re-perform the above operations until the number of iterations meets the termination condition, and select the optimal solution through the hypervolume value (HV) as the final output. Finally, introduce the Pareto elite retention strategy to obtain the optimal solution and the optimal plan.

[0095] To elaborate on the coding method in this paper in detail, a case of scale 10 is used for illustration. In the bilateral local damage disassembly line problem considering electrical restrictions and time-of-use electricity prices, the damage tasks and non-damage tasks are random, and the disassembly time, electricity cost, and energy consumption under different disassembly modes are different. Random coding is likely to generate infeasible solutions, thus reducing the efficiency of the algorithm. Therefore, to ensure the uniqueness of the solution, a double-layer coding based on the task precedence relationship and task attributes is designed. Figure 3 Figure 23 is the precedence relationship diagram of an example product containing 10 disassembly tasks. The numbers in the circles represent the disassembly tasks, and the symbols "h" and "d" represent the hazardous tasks and required tasks respectively. The arrows represent the disassembly precedence relationship of the tasks. According to the task precedence relationship and task attributes, perform double-layer coding. The coding sequence v consists of the disassembly sequence v1 that satisfies the precedence order and the disassembly mode sequence v2, that is, v = [v1: v2]. The specific steps are as follows:

[0096] Step 1: According to Figure 2 in the precedence relationship diagram, the precedence relationship matrix A ij = [a ij n×n can be constructed. If task i is the immediate predecessor task of task j, then a ij = 1, otherwise a ij = 0.

[0097] Step 2: Select the precedence matrix A ij ​​Tasks with a middle column of 0 are used as the disassembly optional task set V. For example, select tasks 1, 4, 5, 9, and 10 to form a disassembly optional task set.

[0098] Step 3: Randomly select a task i in V and put it into the sequence v1, delete the row and column where task i is located, and update the set V;

[0099] Step 4: Randomly select the next disassembly task and put it into the disassembly sequence. Repeat steps 2 - 4 until all disassembly tasks are assigned. The obtained disassembly sequence v1 is the initial feasible solution that conforms to the disassembly precedence relationship, as Figure 2 shown.

[0100] Step 5: Generate the corresponding disassembly mode sequence v2 according to the task sequence v1 in turn. If task i is a dangerous task and a required task, select a non - destructive disassembly mode, and the corresponding v2 is set to 0. As Figure 3 shown, tasks 6, 7, and 9 are required tasks, and task 8 is a dangerous task, so non - destructive disassembly is selected.

[0101] Step 6: if e i > e max , go to step 8, otherwise go to Step 7.

[0102] Step 7: Randomly select a disassembly mode. For example, select destructive disassembly for task 5 and non - destructive disassembly for task 2.

[0103] Step 8: Set the disassembly mode of task i to 1. As Figure 2 shown, tasks 10 and 3 select destructive disassembly.

[0104] Step 9: Repeat steps 5 - 8 until all tasks in the task sequence v1 are assigned disassembly modes.

[0105] Step 10: Output v = [v1:v2].

[0106] After obtaining the disassembly task sequence and the disassembly mode sequence, it is necessary to assign the tasks in the sequence to different workstations under the conditions of meeting the cycle time, disassembly direction, and disassembly mode constraints. This paper adopts a decoding method based on the disassembly mode and disassembly direction. First, determine the disassembly time of the task according to the disassembly mode of the task, and then assign it to the workstation according to the disassembly direction of the task. Under the premise of meeting the beat constraint, assign the tasks without direction constraints according to the principle of the smallest number of workstations.

[0107] S4. Optimizing and updating the initial solution through the Improved Water Cycle Algorithm (IWCA) includes: after generating the initial population after decoding is completed, calculating the objective function, dividing the ocean, rivers, and streams, and generating the first-generation population; subsequently performing ocean-river and river-stream position updates; then performing rainfall operations; combining the offspring population and the parent population to form a new generation of population and repeating the above operations until the number of iterations meets the termination condition, and screening out the optimal solution through the Hypervolume (HV) value as the final output. The specific steps are as follows:

[0108] First, set the problem parameters and algorithm parameters. The problem parameters include the preset beat time CT, disassembly costs (including the electricity costs for performing disassembly tasks in different modes, the electricity costs for turning on workstations, and the additional electricity costs for handling dangerous tasks); the algorithm parameters include the maximum number of iterations (MI), population size (PN), number of rivers (RN), and local variation factor L f ) number of external archives N;

[0109] Second, use the task disassembly sequence obtained through encoding and decoding as the initial individuals to construct the initial population, calculate the objective function, screen out the non-dominated solutions and update the external archive WD, generate the first-generation population, and then determine the ocean, rivers, and streams according to the Pareto elite strategy.

[0110] Subsequently, bilateral crossover is used between the stream and the river to represent the position update between the stream and the river. When the stream flows into the river, two random points r1 and r2 will be generated on the stream. The task sequence between r1 and r2 is used as the crossover segment, and then it is mapped according to the corresponding disassembly task order on the river. The new river segment obtained after mapping will replace the corresponding river segment on the original river segment to form a new river segment. The specific position exchange process between the stream and the river is as Figure 3 shown.

[0111] In addition, three-point crossover is also designed to represent the position exchange between the river and the ocean. Use random numbers to generate three random variants c1, c2, and c3, and divide the mother river and the sea into P1, P2, P3, and P4. Randomly select two river segments as the position information segments. Except for these river segments, the other river segments maintain the sequence order of the parent river. Remove the disassembly tasks from the position information segments of the river, and then copy the task order of the corresponding position information segments in the parent sea area and insert it into the vacant positions of the river to create a new position River new. Similarly, the new sea is generated in the same way. For a detailed description of this operation, please refer to Figure 4 .

[0112] During the rainfall process, new raindrops will form small streams at different positions. Determine the new positions of the newly formed rainflows. To improve the accuracy and prevent convergence to a local optimum. We introduce a new local variation factor L fTo improve the evaporation method and rainfall process of the water cycle. First, generate a random number a. If a < L f , then perform a mutation operation on the sequence. Randomly select a task vk from the sequence v, find the immediate predecessor task vk-1 and the immediate successor task vk+1 of this task, and randomly select a point between the two to insert the task vk, replacing the previous vk position, thereby generating a new individual. As Figure 5 shown.

[0113] Subsequently, combine the updated population with the initial population to form a new generation of population, and re-perform the above operations and update the solution set until the number of iterations meets the termination condition. Finally, screen out the optimal solution through the hypervolume value (HV) as the final output. HV (Hypervolume) refers to the volume of the space formed by the preset reference point distributed in the target space and the Pareto solution set obtained by the algorithm.

[0114] In addition, the HV index, IGD index, and SM index are also introduced to evaluate the convergence and diversity of the algorithm. During the operation of the algorithm, when the HV value tends to be stable, it is considered that the algorithm is in a convergent state. The larger the HV value, the larger the volume of the target space formed by the obtained solutions, indicating that the diversity and quality of the Pareto solution set solved by the algorithm are better. The calculation formula of the HV index is:[[]] In the formula, λ represents the Lebesgue constant to measure the HV value; |S| refers to the number of solutions in the Pareto solution set; v i represents the HV value of the i-th non-dominated solution and the reference point. The inverted generational distance (IGD) index is another key index, which reflects the average distance from each reference point to the ideal Pareto solution. The smaller the IGD value, the better the performance of the algorithm in terms of convergence and distribution. The calculation formula of the IGD index is:[[]] In the formula, P represents the set of points evenly distributed on the actual Pareto surface, and Q represents the optimal Pareto solution set obtained by the algorithm. d(v,Q) refers to the minimum Euclidean distance from the individual v in P to the population Q. In addition, we also introduce the spacing metric (SM), which quantifies the standard deviation of the minimum distance between each solution. The smaller the SM value, the better the uniformity of the solution set distribution. The calculation formula of the SM index is:[[]] In the formula, d i represents the minimum distance between the i-th solution and other solutions in P and all other solutions, and represents the average value of all d i values.

[0115] To further illustrate the effect of the present invention, a specific example is used to illustrate it below.

[0116] The following is the specific program test environment of this embodiment:[[]]

[0117] The simulation computing environment is Intel(R) Core(TM) i5-12400 CPU @ 2.50GHz, 16GB RAM, Windows 11 64-bit, and MATLAB R2020b is used for programming and running.

[0118] To further verify the superiority of the proposed IWCA in solving this problem, four powerful algorithms including Genetic Simulated Annealing Algorithm (GASA), Non-dominated Sorting Genetic Algorithm (NSGA-II), Grey Wolf Optimization Algorithm (GWOA), and Whale Optimization Algorithm (WOA) are selected for comparative testing. Taking a large and complex retired automotive engine as the disassembly case. The 3D drawing and precedence relationship diagram of the retired automotive engine are as Figure 6 shown, which contains a total of 37 disassembly tasks. Figure 6 The names of each part in it are shown in Table 2, and this table also gives information such as the disassembly time, task attributes, electricity cost, power consumption, and the feasibility value of destructive disassembly for each task.

[0119]

[0120]

[0121] Each algorithm runs 10 times and stops after running for 500s each time. The best result of each run is selected for comparison. Since each algorithm obtains dozens of Pareto solutions, 10 solutions are selected using crowding distance, as shown in Table 3. By comparing the results of each group of algorithms together, it is found that the solution set of IWCA completely dominates the other four algorithms, meaning that after the solution space becomes larger as the problem scale increases, the other algorithms can no longer obtain better solutions in a short time. This proves that the invented IWCA has a stronger search ability compared with NSGA-II, GASA, GWOA, and WOA in a short time.

[0122] Table 3 Results of each algorithm for solving the retired automotive engine case

[0123]

[0124] To further test the performance of the algorithm, two multi-objective metrics, namely Inverted Generational Distance (IGD) and Spacing Metric (SM), are introduced and used in combination with HV. Figure 7 The box plots of the HV, IGD, and SM indices of the five algorithms are shown. The higher the HV value, the better the convergence and diversity, while the smaller the IGD and SM values, the better the overall performance and the more uniform the distribution of the solutions. The HV result values after convergence are as Figure 7As shown, i.e., IWCA > GWOA > NSGA-II > WOA > GASA. The HV results show that, compared with the other four algorithms, the proposed IWCA has the best comprehensive performance in solving this case. GWOA is similar to NSGA-II, while the performance of GASA is inferior to the other four algorithms.

[0125] In addition, when evaluating multi-objective algorithms using the IGD and SM metrics, IWCA also demonstrated commendable overall performance and the uniformity of solution distribution. Although the effectiveness of different algorithms may fluctuate depending on the metrics used, IWCA consistently showed strong performance in all three evaluations.

[0126] Considering that the objective preferences of different enterprise managers are different, 4 different solutions were selected from the obtained Pareto solution set. As Figure 8 shown, where red represents destructive tasks and green represents non-destructive tasks. Among these solutions, red represents destructive tasks and green represents non-destructive tasks. Each solution complies with constraints such as task priority, direction, and feasibility of destructive disassembly. The diversity of solutions provides enterprises with the flexibility to make decisions according to their specific goals. For example, if the company prioritizes time efficiency, it may choose S2; if it mainly considers profit, S3 is more preferable; and for environmental protection considerations, S4 will be the best choice.

[0127] As described above, it is only the preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in the embodiments of the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A bilateral local destruction disassembly line setting method considering electrical restrictions and time-of-use electricity prices, characterized in that: It includes the following steps: Step 1: Obtain the basic information of the disassembly task, including the disassembly time under different disassembly modes, the recovery value of the task under different disassembly modes, the unit power consumption of the task under different disassembly modes, the tools required for disassembly, the feasibility value of destructive disassembly, the priority relationship, and the unit power cost at different times of different seasons; Step 2: Establish a multi-objective function for the bilateral local destructive disassembly line, including the number of workstations, the idle time balance rate, the disassembly profit considering time-of-use electricity prices, and the disassembly energy consumption considering different disassembly modes, and establish a corresponding mixed-integer programming model; at the same time, establish the constraint conditions of the model, including the disassembly direction, priority, disassembly mode, time, location, and workstation constraints of the task; Step 3: Solve the mixed-integer programming model; The objective functions of the model constructed in Step 2 are as follows: Minimize the number of workstations: Minimize idle time balancing: Disassembly profit index: Disassembly energy consumption index: In the formula, S k For workstation enable variables, if the kth workstation is enabled, then S k =1, otherwise S k =0; k is the workstation number, CT is the cycle time; i is the disassembly task number; N is the disassembly task number set; l is the left and right index of the workstation, where l=1 means the left side and l=2 means the right side; L is the left and right index set of the workstation; t i represents the disassembly time of task i; G i is the disassembly benefit of the task; E i E is the electricity cost per unit time; k is the fixed unit time cost of the workstation; The additional electricity cost incurred to handle dangerous tasks i; h i is the task hazard attribute variable. If task i is a dangerous task, then h i =1, otherwise h i =0;P i P is the power consumption per unit disassembly time; f is the fixed unit time power consumption of the workstation; P h The power consumption per unit time when processing additional tasks; The variables for assigning tasks to both sides of a workstation. The first bracket of the subscript is the task number, the second bracket of the subscript is the workstation number, and the superscript bracket is the left and right index of the workstation. If the task in the first bracket of the subscript is assigned to the side of the workstation in the first bracket of the subscript, then otherwise t i represents the disassembly time of task i; M is the set of workstations; The constraint conditions of the mixed-integer programming model include: Disassembly direction constraint: D 1 Indicates that the task is the set of tasks on the left, D 2 Indicates that the task is a set of tasks on the right; The relationship between disassembly time, recovery value, electricity cost, and disassembly mode: is the task disassembly mode variable. If task i adopts r disassembly modes, then otherwise r represents the disassembly mode, r = 1 for non-destructive disassembly, r = 2 for destructive disassembly; t i 0 is the disassembly time of task i when non-destructive disassembly is adopted; t i 1 E is the disassembly time when task i is disassembled by destruction; is 0 is the electricity cost per unit time of the processing operation when task i uses non-destructive disassembly in scenario s; E is 1 P is the electricity cost per unit time of the processing operation when task i uses destruction and disassembly in scenario s; i 0 P is the power consumption per unit disassembly time when non-destructive disassembly is used for task i; i 1 The power consumption per unit disassembly time when task i is disassembled by destruction; Disassembly mode constraint: And i >and max ,i∈DP h i +d i ≠0,i∈NP e i is the destructive disassembly feasibility value of task i; e max is the feasibility threshold of destructive disassembly of task i; DP is the set of destructive tasks; h i is the task hazard attribute variable. If task i is a dangerous task, then h i =1, otherwise h i =0;d i is the task requirement attribute variable. If task i is a required task, then d i =1, otherwise d i =0; NP is the set of non-destructive tasks; Workstation constraint: n is the number of disassembly tasks; S k ≤S k-1 ,k∈{2,...,m} m is the number of workstations; is a judgment function. The first bracket of the subscript is the task number of the first task, the second bracket is the task number of the second task, the third bracket is the workstation number, and the superscript bracket is the index of the left and right sides of the workstation. When the first and second tasks are both assigned to the l side of the workstation and the first task is removed before the second task, then otherwise, w (*) is the start disassembly time of a task, and the task number is in the subscript brackets; M is a maximum value; Where A ij =[a ij ] n×n , task i is the immediate predecessor of task j, then a ij =1, otherwise a ij =0; t j represents the disassembly time of task j.

2. The method according to claim 1, characterized in that Step 3 includes: S31: Perform two-stage decoding based on the double-layer coding method, and then generate an initial solution according to the Pareto elite strategy; S32: Optimize and update the initial solution through an improved water cycle algorithm, and introduce the Pareto elite retention strategy to obtain the optimal solution and the optimal plan.

3. The method according to claim 2, characterized in that The optimization and update of the initial solution through the improved water cycle algorithm includes: after decoding is completed, generate an initial population, calculate the objective function, divide the ocean, rivers, and streams, and generate the first-generation population; then perform ocean-river and river-stream position updates; then perform rainfall operations; combine the offspring population and the parent population to form a new generation of population, recalculate the objective function, divide the ocean, rivers, and streams, and generate the second-generation population; then perform ocean-river and river-stream position updates; then perform rainfall operations; repeat the update and rainfall operations until the number of iterations meets the termination conditions, and screen out the optimal solution through the hypervolume value as the final output; Use two-point crossover operation to generate new individuals to replace the individuals in the river for position update; generate new individuals between the river and the ocean through three-point crossover method to replace the original individuals to achieve position update; use the local variation factor Lf to mutate the generated raindrops during the rainfall process: first, generate a random number a, if a < Lf, then perform mutation operation on the sequence, randomly select a task vk from the sequence v, find the immediate predecessor task vk-1 and the immediate successor task vk+1 of this task, and randomly select a point between the two to insert the task vk, replacing the previous vk position, so as to generate a new individual.

Images