Wire oblique cutting blanking method based on parallel grouping genetic algorithm
By optimizing wire beveling using a parallel grouping genetic algorithm, the problem of failing to effectively consider part shape and cutting process in existing technologies is solved, resulting in more efficient wire utilization and intelligent management, and reduced costs.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-01
- Publication Date
- 2026-03-31
AI Technical Summary
Existing wire cutting methods fail to effectively consider the shape of parts and cutting processes, resulting in waste and increased costs, and lack of intelligent management.
Parallel grouping genetic algorithm is used to optimize wire oblique cutting. By improving the FF algorithm and BFD algorithm, and combining elite selection, Exon crossover and adaptive mutation, the wire cutting scheme is optimized to reduce cutting loss and raw material consumption.
It improved wire cutting efficiency, reduced waste, lowered raw material costs, and enabled intelligent production management, thereby enhancing the company's production efficiency and economic benefits.
Smart Images
Figure CN115422823B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the technical field of wire EDM blanking, and more specifically, relates to a wire beveling blanking method based on a parallel grouping genetic algorithm. Background Technology
[0002] For wire processing companies that cut and manufacture parts such as pipes and profiles, cost reduction and efficiency improvement are of great significance. Adopting better wire cutting methods can not only reduce waste and lower the cost of raw materials, but also, based on computer technology, intelligent and information-based cutting can reduce the time cost of manual cutting and facilitate the company's wire production management.
[0003] By reviewing relevant literature, it was found that existing wire cutting methods simplify the wire into a one-dimensional wire model with only length and no cross-section, and apply one-dimensional cutting mathematical models and algorithms to solve the problem. However, they do not take into account the cutting loss of wire parts and the rotation of parts, which are related to the shape of wire parts and the cutting process, and directly affect the cutting effect and cost. Summary of the Invention
[0004] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a wire beveling method based on a parallel grouping genetic algorithm. This method optimizes the wire beveling process using a parallel grouping genetic algorithm to find the cutting scheme that consumes the fewest raw materials, thereby improving the cutting efficiency and reducing the raw material cost of wire parts.
[0005] To achieve the above objectives, according to one aspect of the present invention, a method for oblique cutting of wire based on a parallel grouping genetic algorithm is provided, the method mainly comprising the following steps:
[0006] Step 1: Generate m subpopulations P of size n. i0 The overall population P0 is formed. The maximum number of iterations of the algorithm is set to G1, the number of population drifts is set to G2, and the number of iterations is set to g. The group coding sequence of each individual is randomly initialized, and then the improved FF algorithm is used to decode and feed the data and calculate the fitness value of the individual.
[0007] Step 2, for each subpopulation P ig Individuals are sorted in descending order of fitness. The top e elite individuals (i.e., those with higher fitness) are then selected to replace the remaining n poorer individuals, and all of them are then replicated to the next generation subpopulation P. ig+1 middle;
[0008] Step 3: Select each subpopulation P in turn. ig If two adjacent individuals satisfy the crossover condition, an Exon crossover is performed to generate a new individual. The fitness of each individual is calculated, and the two individuals with the larger fitness are copied to the new subpopulation P.ig+1 In the process, all subpopulations are traversed to form a new subpopulation, and then individuals from each subpopulation are arranged in descending order of fitness.
[0009] Step 4: Select each subpopulation P in turn. ig If an individual in the population meets the mutation condition, several mutation sites are selected for mutation rearrangement, fitness is calculated and replicated into a new population. After traversing all subpopulations to form a new subpopulation, individuals from each population are arranged in descending order of fitness.
[0010] Step 5, iteration number g = g + 1, check if g is less than G2. If so, perform population drift, i.e., use each subpopulation P ig The best individual in the population replaces the next a subpopulations P. ag For each individual, calculate the fitness value of the last individual, the second to last individual, and so on up to the ath individual from the end, and sort them in descending order of their size; otherwise, go directly to Step 6.
[0011] Step 6: Determine if g is greater than G1. If so, decode according to the group encoding and output the final material cutting scheme; otherwise, go to Step 2.
[0012] Furthermore, the individual grouping coding divides the chromosome of each individual in the subpopulation into a target part and a subgroup part. The target part uses decimal integer coding to represent the numbers of q parts, and the subgroup part uses English letters to represent the numbers of k raw materials. In the coding, the wire parts are divided into different subgroups, and each subgroup corresponds to a raw material, indicating that these parts are cut on this raw material. The individual coding sequence between [1, q] is randomly initialized and then converted into group coding using the FF algorithm. This operation is performed on each individual to complete the initialization of the entire population.
[0013] Furthermore, converting it into block coding using the FF algorithm includes the following steps:
[0014] Step 1.1: Select the parts with the corresponding numbers according to the sequence codes, and place the i-th part into the first column that satisfies the remaining length L. r Find the raw material S(i-1,i) and update the remaining length L of the raw material. r If no suitable material is found, add a new material and insert it into the part, then update the remaining material length L. r ;
[0015] Step 1.2: Combine the part number and the corresponding raw material number to form a group code;
[0016] Step 1.3: Repeat the above steps until all parts have been traversed.
[0017] Furthermore, the crossover condition refers to generating a random number r within the range [0,1]. c If r c <P c If the crossover probability is specified, then the two selected individuals will be crossovered; otherwise, they will not be crossovered.
[0018] Furthermore, the crossover probability P c The adaptive probability formula is used to calculate the probability. The adaptive probability calculation formula is as follows:
[0019]
[0020] Where P c P represents the adaptive crossover probability. c1 and P c2 Let f represent the maximum and minimum crossover probabilities, and let f represent the individual fitness. avg f represents the average fitness in the subpopulation. max f' represents the maximum fitness in the subpopulation, and f' represents the larger fitness among the two individuals.
[0021] Furthermore, the Exon crossover method includes the following steps:
[0022] Step 3.1: Merge the two selected individuals and sort them in ascending order of the remaining length of each raw material. If the remaining lengths of the raw materials are the same, sort them in ascending order of the number of parts cut from each raw material. If the number of parts is still the same, sort them according to the size of the raw material cutting number.
[0023] Step 3.2: Copy all the blanking parts and arrange them in descending order of the sum of their straight length and the lengths of their two diagonal angles. Check each blanking material of the merged individual in turn. If the part on the current material can be found in the copied parts, then place the copy part corresponding to this material into the new material as a blanking material of the offspring. Continue until all the merged blanking materials have been traversed. If there are still some parts that have not been blanked, then place these parts into materials with sufficient remaining length according to the improved BFD algorithm.
[0024] Furthermore, the mutation condition refers to generating a random number r between [0,1]. m If r m <P m If the selected individual is mutated, then the mutation operation is performed; otherwise, no mutation is performed. The specific mutation method is based on the selection probability P of the mutation site. b If a random number r between [0,1] is generated b Satisfying r b <P bIf the result is mutated, it means that the root material is mutated, all parts on the selected material are removed, and then the materials are re-arranged according to the improved BFD algorithm.
[0025] Furthermore, the mutation probability P m and site selection probability P b The calculation formulas are as follows:
[0026]
[0027]
[0028] Furthermore, the local part rearrangement in the BFD algorithm includes the following steps:
[0029] Step 1.1: Arrange all the parts to be cut in descending order of the sum of the straight section length and the length of the bevel at both ends;
[0030] Step 1.2: Place the i-th part into the first material that satisfies the condition that the remaining length Lr > S(i-1,i), and update the remaining length Lr of the material; if no material can be placed into the part, add a new material and place the part into it, and update the remaining length Lr of the material.
[0031] Step 1.3: Sort Lr in ascending order, and arrange the raw materials in ascending order of Lr;
[0032] Step 1.4: Combine the part number and the corresponding raw material number to form a group code;
[0033] Step 1.5: Repeat all the above steps until all parts have been traversed.
[0034] Furthermore, fitness refers to the following objective function and constraints, where,
[0035] Objective function:
[0036]
[0037] Where f represents the optimization objective; F is the fitness value of each material cutting scheme; l j This indicates the length of wire used for cutting the j-th type of blanking method; L represents the length of the raw material being cut; x j This indicates the quantity of the j-th material feeding method used;
[0038] The constraints are:
[0039]
[0040]
[0041] Among them, aij d represents the quantity of part j to be cut in each cutting method i. i Indicates the required quantity of each type of part, l j The length l of the straight section of the wire component i Cut-off matching loss c ij The losses from cutting the parts at the beginning and end of the raw material must not exceed the length L of the raw material.
[0042] In summary, compared with the prior art, the wire beveling method based on parallel grouping genetic algorithm provided by this invention has the following advantages:
[0043] 1. The method of this invention is based on parallel grouping genetic algorithm for wire cutting optimization. It takes into account the wire cutting loss. By improving the FF algorithm, the coding sequence is transformed into group coding to represent a cutting scheme. The coding is intuitive and facilitates the crossover and mutation operations of the algorithm. It is faster and has better optimization effect than conventional coding search. By improving the FF algorithm and the BFD algorithm, that is, rotating when inserting wire parts, cutting loss is minimized. The determination of whether a part can be inserted is also based on comparing the remaining length and the equivalent length of the part, so that wire oblique cutting can be achieved.
[0044] 2. A parallel genetic algorithm framework and population migration strategy are adopted, which increases the population search space through parallel search of multiple subpopulations and the exchange of superior individuals. An elite selection strategy and the Exon crossover operator are used to fully retain excellent individuals while improving the efficiency of crossover search. An adaptive crossover and mutation strategy is adopted, which adaptively adjusts the crossover and mutation probabilities as the fitness value of individuals changes with algorithm iteration. The algorithm converges faster in the early stage of iteration and has a greater crossover and mutation probability when retaining the best individuals in the later stage, thereby improving the problem of easily getting trapped in local optima. The selection of mutation sites is also based on the utilization rate of each raw material, which can reduce the damage to the better raw materials and further improve the convergence performance.
[0045] 3. By utilizing the method of this invention, enterprises can not only have better material cutting methods, reduce waste caused by scrap materials and save wire raw materials; but also achieve intelligent and information-based linkage of supply, inventory, material cutting and outbound based on computer technology, reducing the time cost of manual sampling, and facilitating the production management of enterprises. Attached Figure Description
[0046] Figure 1 This is a flowchart illustrating a wire beveling method based on a parallel grouping genetic algorithm provided by the present invention. Detailed Implementation
[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.
[0048] Please see Figure 1 This invention provides a wire beveling method based on a parallel grouping genetic algorithm. The method employs a parallel grouping genetic algorithm to solve the problem, using an elite selection strategy to retain the best solutions in the population. Based on the problem characteristics, a grouping encoding method is used, with one encoding sequence representing one solution. Considering wire cut matching, an improved BFD algorithm is proposed, transforming the encoding sequence into a wire part cutting scheme on the raw material and calculating the fitness value of each individual. An Exon crossover method is used for adaptive crossover operations. Adaptive mutation operations are performed using adaptive probability and novel mutation site selection. In the initial few iterations, a population drift operation is used to exchange superior individuals among the subpopulations. The wire beveling method proposed in this invention can effectively solve the practical wire beveling optimization problem, improve wire cutting efficiency and material utilization, and generate significant economic benefits.
[0049] The material feeding method mainly includes the following steps:
[0050] Step 1: Initialize algorithm-related parameters and generate m subpopulations P of size n. i0 The overall population P0 is formed. The maximum number of iterations of the algorithm is set to G1, the number of population drifts is set to G2, and the number of iterations is set to g. The group coding sequence of each individual is randomly initialized, and then the improved FF algorithm is used to decode and cut the material and calculate the fitness value of the individual. The objective function is: the minimum number of raw material pieces consumed in the cutting.
[0051] Individual grouping encoding divides the chromosome of each individual in the subpopulation into a target part and a subgroup part. The target part uses decimal integer encoding to represent the numbers of q parts, and the subgroup part uses English letters to represent the numbers of k raw materials for cutting. In the encoding, wire parts are divided into different subgroups, and each subgroup corresponds to a raw material for cutting, indicating that these parts are cut from this raw material. A random initialization generates an individual encoding sequence between [1, q], and then the improved FF algorithm is used to convert it into a group encoding. This operation is performed on each individual to complete the initialization of the entire population.
[0052] The improved FF algorithm for generating block codes includes the following steps:
[0053] Step 1.1: Select the parts with the corresponding numbers according to the sequence codes, and place the i-th part into the first column to satisfy the remaining length L.r Find the raw material S(i-1,i) and update the remaining length L of the raw material. r If no suitable material is found, add a new material and insert it into the part, then update the remaining material length L. r ;
[0054] Step 1.2: Combine the part number and the corresponding raw material number to form a group code.
[0055] Step 1.3: Repeat the above steps until all parts have been traversed.
[0056] Step 2: Perform the elite selection operation, and select each subpopulation P ig Individuals are sorted in descending order of fitness. The top e elite individuals (i.e., those with higher fitness) are then selected to replace the remaining n poorer individuals, and all of them are then replicated to the next generation subpopulation P. ig+1 middle.
[0057] Step 3: Select each subpopulation P in turn. ig If two adjacent individuals satisfy the crossover condition, an Exon crossover is performed to generate a new individual. The fitness of each individual is calculated, and the two individuals with the larger fitness are copied to the new subpopulation P. ig+1 In the process, all subpopulations are traversed to form a new subpopulation, and then individuals from each subpopulation are arranged in descending order of fitness.
[0058] Crossover condition refers to generating a random number r within the range [0,1]. c If r c <P c If the crossover probability is P, then the two selected individuals undergo a crossover operation; otherwise, no crossover occurs. c The adaptive probability formula is used to calculate the probability. The adaptive probability calculation formula is as follows:
[0059]
[0060] Where P c P represents the adaptive crossover probability. c1 and P c2 Let f represent the maximum and minimum crossover probabilities, and let f represent the individual fitness. avg f represents the average fitness in the subpopulation. max f' represents the maximum fitness in the subpopulation, and f' represents the larger fitness among the two individuals.
[0061] The Exon crossover method includes the following steps:
[0062] Step 3.1: Merge the two selected individuals and sort them in ascending order of the remaining length of each raw material. If the remaining lengths of the raw materials are the same, sort them in ascending order of the number of parts cut from each raw material. If the number of parts is still the same, sort them according to the size of the raw material cutting number.
[0063] Step 3.2: Copy all blanking parts and arrange them in descending order of the sum of their straight length and the lengths of their two diagonal angles. Check each blanking material of the merged individual in turn. If the part on the current material can be found in the copied parts, then place the copy part corresponding to this material into the new material as a blanking material of the offspring. Continue until all the merged blanking materials have been traversed. If there are still some parts that have not been blanked, then place these parts into materials with sufficient remaining length according to the improved BFD algorithm.
[0064] Step 4: Select each subpopulation P in turn. ig If an individual in the population meets the mutation condition, several mutation sites are selected for mutation rearrangement, fitness is calculated and replicated into a new population. This process is repeated for all subpopulations to form a new subpopulation, and then individuals from each population are arranged in descending order of fitness.
[0065] The mutation condition refers to generating a random number r between [0,1]. m If r m <P m If the mutation probability is P, then the selected individual undergoes mutation; otherwise, no mutation occurs. The specific mutation method is based on the mutation site selection probability P. b If a random number r between [0,1] is generated b Satisfying r b <P b If the result is positive, it indicates that the root material undergoes mutation. All parts on the selected root material are removed, and then the root materials are re-arranged according to the improved BFD algorithm; where the mutation probability P is... m and site selection probability P b The calculation formulas are as follows:
[0066]
[0067]
[0068] The improved BFD algorithm for local part rearrangement includes the following steps:
[0069] Step 1.1: Arrange all parts to be cut in descending order of the sum of the straight section length and the bevel length at both ends.
[0070] Step 1.2: Place the i-th part into the first material that satisfies the remaining length Lr > S(i-1, i), and update the remaining length Lr of the material; if no material can be placed, add a new material and place the part into it, and update the remaining length Lr of the material.
[0071] Step 1.3: Sort Lr in ascending order, and arrange the raw materials in ascending order of Lr;
[0072] Step 1.4: Combine the part number and the corresponding raw material number to form a group code.
[0073] Step 1.5: Repeat all the above steps until all parts have been traversed.
[0074] Step 5: After iteration g = g + 1, check if g is less than G2. If so, perform population drift, i.e., use each subpopulation P ig The best individual in the population replaces the next a subpopulations P. ag Calculate the fitness value of each individual from the last one to the last a, and sort them in descending order of their size; otherwise, proceed directly to Step 6.
[0075] The fitness mentioned in Steps 1 through 5 refers to the following objective function and constraints:
[0076] Objective function:
[0077]
[0078] Where f represents the optimization objective; F is the fitness value of each material cutting scheme; l j This indicates the length of wire used for cutting the j-th type of blanking method; L represents the length of the raw material being cut; x j This indicates the number of times the j-th material feeding method is used.
[0079] Constraints:
[0080]
[0081]
[0082] Among them, a ij d represents the quantity of part j to be cut in each cutting method i. i Indicates the required quantity of each type of part, l j The length l of the straight section of the wire component i Cut-off matching loss c ij The losses from cutting the parts at the beginning and end of the raw material must not exceed the length L of the raw material.
[0083] Step 6: Update and save the best individual, check if g is greater than G1. If so, decode according to the grouping code and output the final material cutting scheme. Otherwise, go to Step 2 to continue searching.
[0084] The present invention will be further described in detail below with reference to specific embodiments.
[0085] For ease of explanation, the following symbols introduced in this embodiment are explained in the table below:
[0086] Symbol definition
[0087]
[0088] The preferred embodiment of the present invention addresses the problem of wire beveling and blanking considering slit losses, a special type of one-dimensional blanking problem. Specifically, this problem studies the cutting of n different sizes, each with a quantity of d. i The wire parts are cut from raw materials of length L with a sufficient supply. The cutting sequence and cutting method are optimized to minimize the number of raw material pieces k consumed.
[0089] Optimization Model
[0090] The wire beveling problem solved in this embodiment refers to cutting m different sizes of wire, with each size having a quantity of d. i The wire parts are cut from raw materials of length L with a sufficient supply. The cutting sequence and cutting method are optimized to minimize the number of raw material pieces k consumed.
[0091] The objective function in this embodiment is as follows:
[0092]
[0093] This formula represents the optimization objective for the problem of wire beveling.
[0094] The constraints are as follows:
[0095]
[0096]
[0097] These two formulas respectively indicate that the number of parts to be cut is equal to the number of parts required for material cutting, and that the total length of the parts to be cut cannot exceed the length of the original material.
[0098] The objective function and constraints described above are a preferred choice for minimizing the number of raw material roots k proposed in this embodiment. In other embodiments, constraints can be added or deleted, and other objective functions and constraints for evaluating the sum of values can be formulated according to actual needs, and are not limited to this embodiment.
[0099] Encoding and Decoding
[0100] Since the problem solved by the method provided in this embodiment is a grouping and combination optimization problem, a grouping coding method is adopted according to the characteristics of the problem. That is, each wire part number is represented by a decimal integer (the value range is [1, q]), forming a decimal sequence. The position in the code is the order in which the wire part is placed into the raw material. The decimal integer of the target part represents the raw material number. The two sets of codes together form the grouping code of the part and its raw material.
[0101] Decoding involves converting the decimal block encoding sequence into a specific material cutting scheme and then calculating the total value, as shown in Table 3. The block encoding method used in this embodiment can intuitively represent a material cutting scheme without the need for additional decoding operations.
[0102] Table 3. Block coding based on the improved FF algorithm
[0103]
[0104] Implementation Cases
[0105] The following detailed description of this embodiment is based on Table 3, using a square tube raw material with a length L = 3000 mm, cross-sectional dimensions W = 100 mm, H = 100 mm, and wall thickness E = 3 mm, and five different lengths of square tube parts with the same cross-section to be cut as examples.
[0106] Table 1. Example of square tube beveling and blanking: SX1 part dimensions and quantity.
[0107]
[0108] According to the data in Table 1, the number of parts in this example is q = 22. Therefore, a random sequence [1, q] is generated, assuming S = [22 6 3 16 11 7 17 14 8 5 2 1 19 15 1 2 4 18 13 9 20 10 12]. The square tube parts are arranged in this numbering order using the improved FF algorithm to form the cutting scheme shown in Table 4. In Table 4, the top row of the numbering sequence represents the part numbers, and the code below represents the raw material numbers, together forming the grouping code of the cutting scheme. The crossover and mutation process involves rearranging the part numbers corresponding to the raw material numbers using the improved BFD algorithm according to the Exon crossover and grouping mutation method. Finally, through iterative evolution, the optimal cutting result shown in Table 2 can be obtained.
[0109] Table 2. Solution results for the example of oblique cutting of square tubes.
[0110]
[0111]
[0112] Table 2 shows the results of a square tube beveling calculation obtained using this method. It includes three cutting methods, each with its own quantity, part number, and rotation angle for each part to be inserted into the raw material. It also includes the remaining material length and utilization rate for each method, which is very convenient for those skilled in the art to understand and apply.
[0113] Table 4. Solution results for the oblique cutting example.
[0114]
[0115] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements 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 wire beveling blanking method based on a parallel grouping genetic algorithm, characterized by, The method comprises the following steps: Step1, generating m a sub-population of size n P i0 constituting the whole population P 0, set the maximum iteration number of the algorithm as G 1, the population drift number as G 2, the iteration number as g ; randomly initialize the grouping coding sequence of each individual, then decode the blanking by using the improved FF algorithm and calculate the fitness value of the individual; Step 2, arrange the individuals in each sub-population in descending order of fitness, then select the first individuals with higher fitness to replace the last individuals with lower fitness, and copy them to the next generation sub-population together. e n-e Step 3, select each sub-population in turn P ig If the two adjacent individuals satisfy the crossover condition, a new individual is generated by using Exon crossover, the fitness is calculated, and the larger of the two individuals is copied to the new sub-population P ig+1 In the middle, all sub-populations are traversed to form a new sub-population, and the individuals of various populations are arranged in descending order according to the fitness Step4, select each sub-population in turn P ig If the mutation condition is met, select several mutation sites for mutation rearrangement, calculate the fitness and copy to the new population. After traversing all sub-populations to form a new sub-population, arrange the individuals in the population in descending order of fitness. Step 5, Number of iterations g = g +1, judgment g Is it less than G 2. If so, then population drift is performed, i.e., each subpopulation is used. P ig The optimal individual in the sequence, respectively replacing the subsequent ones. a Subspecies P ag The last one, the second to last, and so on up to the last one. a For each individual, calculate the fitness value of each individual and sort them in descending order of size; otherwise, proceed directly to Step 6. Step 6, judging g whether it is greater than G 1, if yes, decoding according to the grouping coding, outputting the final blanking scheme, otherwise, going to Step 2; The fitness refers to the following objective function and constraint conditions, wherein, The objective function is: in, Indicate the optimization objective; F It is the fitness value for each material cutting scheme; It means the first j The length of wire parts used for cutting in the seed material preparation method; L Indicates the length of the raw material to be cut; Indicates the first j The quantity of materials used in the seeding and feeding method; The constraint condition is: wherein, represents the number of parts i in each blanking mode j , represents the required number of each part, consisting of the straight length of the wire part l i , the loss of cut matching c ij and the loss of the first and last cutting of the part on the raw material, must satisfy the condition that it does not exceed the length of the raw material L .
2. The parallel grouping genetic algorithm based wire beveling blanking method of claim 1, wherein: The group coding of individuals divides the chromosomes of each individual in the sub-population into target parts and group parts, the target parts are expressed by decimal integer coding q The part number, the group parts are expressed by English letter coding k The original material number of the blank, in the coding, the wire parts are divided into different groups, each group corresponds to an original material of the blank, indicating that these parts are cut on this original material; randomly initialize to generate [1, q ] The individual coding sequence between them, and then convert it into group coding by using the FF algorithm. This operation is performed on each individual to complete the initialization of the entire population.
3. The parallel grouping genetic algorithm based wire beveling blanking method of claim 2, wherein: The FF algorithm is used to convert the grouping code, which comprises the following steps: Step 1.1, according to the sequence coding, select the corresponding numbered parts, the first i part into the first raw material that meets the remaining length L r S i -1, i ) and update the remaining length of the raw material L r ; if no raw material can be found, add a new raw material and arrange the part, update the remaining length of the raw material L r ; Step 1.2, combining the part number and the corresponding raw material number to form a grouping code; Step 1.3, repeating the above steps until all parts are traversed.
4. The parallel grouping genetic algorithm based wire beveling blanking method of claim 1, wherein: The crossover condition is to generate a random number in [0,1] r c If r c P c Two selected individuals are crossed, otherwise not. 5. The parallel grouping genetic algorithm based wire beveling blanking method of claim 4, wherein: Cross probability P c is calculated using an adaptive probability formula, which is: wherein P c denotes the adaptive crossover probability, P c1 and P c2 denotes the maximum and minimum crossover probability, f denotes the individual fitness, f avg denotes the average fitness in a subpopulation, f max denotes the maximum fitness in a subpopulation, f’ denotes the greater fitness of two individuals.
6. The wire beveling blanking method based on parallel grouping genetic algorithm according to claim 5, characterized in that: The exon cross mode comprises the following steps: Step 3.1, merging the two selected individuals, sorting according to the remaining length of each raw material from small to large, if the raw material remaining length is the same, then sorting according to the number of parts cut on each raw material from small to large, if the number of parts is still the same, then sorting according to the raw material blanking number size; Step 3.2, copying all the blanking parts, arranging them in the order of the sum of the straight length and the length of the two end bevels from large to small, checking each blanking raw material of the merged individual in turn, if the parts on the current raw material can be found from the copied parts, then the copied parts corresponding to the raw material are arranged into a new raw material as a blanking raw material of the offspring, until all the blanking raw materials of the merged individual are traversed; if there are still parts left that are not used by the blanking, then these parts are arranged into the raw material with sufficient remaining length according to the improved BFD algorithm.
7. The parallel grouping genetic algorithm based wire beveling blanking method of claim 1, wherein: The mutation condition refers to generating a random number between [0,1]. r m ,like r m < P m If the selected individual is mutated, then the mutation operation is performed; otherwise, no mutation is performed. The specific mutation method is based on the probability of selecting the mutation site. P b If a random number between [0, 1] is generated r b ,satisfy r b < P b If the result is mutated, it means that the corresponding raw material is mutated, all parts on the selected raw material are removed, and then the raw materials are re-arranged according to the improved BFD algorithm.
8. The parallel grouping genetic algorithm based wire beveling blanking method of claim 7, wherein: Mutation probability P m and site selection probability P b The calculation formulas are as follows: In the formula, f avg represents the average fitness in the subpopulation.
9. The parallel grouping genetic algorithm based wire beveling blanking method of claim 8, wherein: The BFD algorithm part local rearrangement comprises the following steps: Step 1.1, arranging all the parts to be blanked in the order of the sum of the straight length and the length of the two end bevels from large to small; Step 1.2, the first i The parts are arranged into the first column to satisfy the remaining length. Lr > S ( i -1, i The raw materials are updated, and the remaining length of the raw materials is updated. Lr If no suitable material is found, add a new material and insert it into the part, then update the remaining length of the material. Lr ; Step 1.3, to Lr In ascending order, and the raw materials according to Lr In ascending order; Step 1.4, combining the part number and the corresponding raw material number to form a grouping code; Step 1.5, repeating all the above steps until all parts are traversed.