A nesting optimization method and system based on the best matching rule and tabu search

Abstract

The present invention provides a layout optimization method and system based on the best matching rule and tabu search, belonging to the technical field of layout optimization. The characteristic of backward influence existing in the layout process causes the inability to correspond the adjustment in sequence with the adjustment in layout, restricting the effectiveness of sequence adjustment. Traditional methods do not consider this characteristic and adopt unnecessary and complex inefficient operations. On the basis of retaining the high positioning efficiency of the best matching rule, the present invention changes the way of sequence update, selects a certain part in the original sequence as the demarcation point for retention and change in layout, and avoids complex operations in sequence adjustment, but only adopts the exchange operation between this part and the subsequent parts of this part. In this way, the layout of the previous parts of this part is retained, and the layout of the subsequent parts of this part is changed, which is more in line with the way of part layout update, avoids excessive unnecessary and inefficient complex operations adopted by traditional methods, and improves the search efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of nesting optimization, and more specifically, relates to a nesting optimization method and system based on the best matching rule and tabu search. Background Art

[0002] In the digital manufacturing process of the industrial cutting industry, the nesting optimization problem exists in the design link of the blanking and cutting link, that is, to reasonably design the layout of parts with predetermined shapes on a given template, and the parts cannot overlap and are closely abutted, so as to achieve the purpose of improving the material utilization rate. The nesting of rectangular parts is a branch problem in the nesting optimization problem, in which the shapes of the parts are all rectangles, and the parts are placed orthogonally, that is, the boundaries of the rectangular parts are parallel or perpendicular to the boundaries of the plate, and this method can ensure the tightness of the part arrangement.

[0003] The solution method of the nesting optimization problem is called the nesting optimization algorithm. A typical nesting optimization algorithm consists of two parts: a positioning strategy and a search algorithm. The positioning strategy is to determine the position where the part should be placed by setting heuristic positioning rules, and the search algorithm is to change the order and angle of the parts on the basis of the positioning strategy, construct different nesting layouts and select high-quality layouts from them. Commonly used positioning strategies include the bottom left rule and the best fit rule. Both the bottom left rule and the best fit rule require setting the order of the parts, but the bottom left rule performs nesting according to the set part order, placing each part at the bottom leftmost position on the plate, while the best fit rule does not perform nesting completely according to the set part order, but traverses the remaining parts and selects the most suitable part in the current nesting situation and places it at the selected position on the plate.

[0004] Since the parts are placed at the selected positions on the plate one by one in sequence during the nesting process, each placed part will change the existing part boundary line structure, and the boundary line structure determines the selection of the next part and the placement position of the part. Therefore, the nesting process has the characteristic of backward influence, that is, the parts placed first always affect the nesting process of the parts placed later.

[0005] Search algorithms usually aim at gradually optimizing the nesting quality. By means of iteration, the order of parts is adjusted. The adjustment method is to keep the order of some parts and change the order of the rest. When executing the search algorithm, the new order of parts is generated by adjusting the original order, and both the original order and the new order need to form a nesting layout by placing parts one by one. Due to the characteristic of backward influence in the nesting process, the first part whose order changes will change the part boundary line structure in the nesting layout corresponding to the original order, that is, starting from this part, the layout of all subsequent parts will change. Therefore, the adjustment of the order cannot be transformed into the corresponding adjustment of the layout, and the search algorithm still has the defect of insufficient search efficiency. CN103500255A discloses an intelligent nesting method for rectangular parts, whose rectangular nesting meets the one-cut blanking process, adopts the positioning strategy of the lowest horizontal line, and combines the genetic algorithm to search for the order of parts. CN113032921A discloses a nesting algorithm based on parallel adaptive parameter cuckoo search and the lowest horizontal line, which also adopts the positioning strategy of the lowest horizontal line, improves the cuckoo algorithm, and enhances the robustness of the search through the parallel strategy and the parameter adaptive strategy. CN110909947A discloses a nesting method and device for rectangular parts based on the grey wolf algorithm, which also adopts the positioning strategy of the lowest horizontal line, improves the encirclement, hunting and wandering of the traditional grey wolf algorithm, simplifies the algorithm process and enhances the search ability at the same time. These methods all adopt the lowest horizontal line positioning rule belonging to the best matching rule, but the search algorithms do not consider the characteristic of backward influence in the nesting process, so the search efficiency still needs to be improved. Summary of the Invention

[0006] In view of the above defects or improvement requirements of the prior art, the present invention provides a nesting optimization method and system based on the best matching rule and tabu search, aiming to solve the technical problem that the existing search algorithm takes complex and unnecessary sorting operations due to not considering the characteristic of backward influence in the nesting process, resulting in insufficient search efficiency.

[0007] To achieve the above object, according to one aspect of the present invention, there is provided a nesting optimization method based on the best matching rule and tabu search, including:

[0008] S1. Randomly initialize the order sequence and orientation sequence of rectangular parts;

[0009] S2. Execute a rectangular nesting algorithm based on the best matching rule on the initialized rectangular parts to obtain a nesting information solution X; the nesting information solution includes an actual order sequence, an actual orientation sequence, a shape matching degree sequence composed of shape matching degree values corresponding to each part, and the layout occupation length obtained from this nesting.

[0010] S3. Select a part with a shape matching degree value of 0 from the nesting information solution X as the part to be adjusted p;

[0011] S4. Exchange the part located after the part to be adjusted p and having a different dimension in the x direction from p with p, regenerate the actual part sequence, actual orientation sequence, shape matching degree sequence, and layout occupied length, and construct a dynamic neighborhood solution set;

[0012] S5. Select a solution with the shortest layout occupied length and different from the nesting information solution corresponding to the historical iteration round from the dynamic neighborhood solution set as the new nesting information solution;

[0013] S6. Determine whether the set number of loops is reached; if not, return to execute step S3; if so, use the layout corresponding to the current nesting information solution as the final nesting result.

[0014] Furthermore, the nesting algorithm based on the best matching rule specifically includes the following process:

[0015] Select the gap at the bottom leftmost position as the position for placing the part, and use the bottom left docking positioning;

[0016] Traverse all the remaining parts, and find the part with the best shape match with the selected gap as the next nesting part; if there are multiple parts with the maximum shape matching degree at the same time, select the part that is the most forward in the sequence.

[0017] Furthermore, the judgment criterion for the shape match with the selected gap is whether the part can completely fit the boundary of the selected gap; if the part does not have a boundary fit with the selected gap, the shape matching degree is 0; if the part has a boundary fit with the selected gap for a certain boundary, the shape matching degree is 1; if the part has a boundary fit with the selected gap for two certain boundaries, the shape matching degree is 2; if the part has a boundary fit with the selected gap for all three boundaries, the shape matching degree is 3.

[0018] Furthermore, the specific process of performing the exchange operation in step S4 is,

[0019] 01. Judge whether the actual occupied length of the part q located after the part to be adjusted p in the x direction when placed at the original angle is the same as the actual occupied length of the part p in the x direction; if the same, perform the exchange operation; otherwise, go to 02;

[0020] 02. Judge whether the actual occupied length of the part q when rotated by 90 degrees in the x direction is the same as the actual occupied length of the part p in the x direction; if the same, perform the exchange operation; if not, sequentially select the next part and go to 01.

[0021] Corresponding to the execution process of each step of the above method, the present invention also provides a nesting optimization system based on the best matching rule and tabu search, including:

[0022] Initialization module, randomly initialize the sequence of the rectangular parts and the sequence of orientations.

[0023] Initial nesting module, perform the rectangular nesting algorithm based on the best matching rule on the initialized rectangular parts to obtain the nesting information solution X; the nesting information solution includes the actual sequence, the actual orientation sequence, the shape matching degree sequence composed of the shape matching degree values corresponding to each part, and the layout occupied length obtained from this nesting.

[0024] Shape matching degree screening module, select a part with a shape matching degree value of 0 from the nesting information solution X as the part to be adjusted p.

[0025] Dynamic neighborhood solution set construction module, exchange the parts that are located after the part to be adjusted p and have different dimensions in the x direction with p, regenerate the actual sequence of parts, the actual orientation sequence, the shape matching degree sequence, and the layout occupied length, and construct a dynamic neighborhood solution set.

[0026] Tabu search module, select the solution with the shortest layout occupied length and different from the nesting information solution corresponding to the historical iteration round from the dynamic neighborhood solution set as the new nesting information solution.

[0027] Loop judgment module, judge whether the set number of loops is reached; if not, return to the shape matching degree screening module; if so, take the layout corresponding to the current nesting information solution as the final nesting result.

[0028] Generally speaking, compared with the prior art by the above technical solutions conceived by the present invention, the following beneficial effects can be achieved.

[0029] (1) Traditional methods represented by genetic algorithms rely on operations such as crossover and mutation to generate new sequences from the original sequence in sequential search. The operations adopted are usually complex operations composed of operations such as exchange, insertion, and rotation. Due to the characteristic of backward influence in the nesting process, the adjustment of the sequence cannot correspond to the adjustment of the layout. The complex operations adopted by traditional methods on the part sequence are unnecessary and inefficient operations, so the search efficiency is insufficient. Compared with traditional methods, this method changes the way of sequence update, selects a certain part in the original sequence as the demarcation point for retention and change in the layout, and avoids complex operations on sequence adjustment, but only adopts the exchange operation between this part and the parts after this part. In this way, the layout of the parts before this part is retained, and the layout of the parts after this part is changed, avoiding excessive unnecessary and inefficient complex operations adopted by traditional methods and improving the search efficiency.

[0030] (2) In the process of selecting a certain part as the demarcation point for retention and change in the layout from the original order in this method, the judgment of the shape matching degree value of the part is combined, and only the part with the shape matching degree value of 0 is selected, that is, only the part that cannot match the shape of the boundary line is selected as the adjustment object, fully retaining the high positioning efficiency of the best matching rule.

[0031] (3) In the swap operation adopted in this method, a part with a different size in the x direction from the demarcation point part is selected as the swap object to ensure that the layout of the subsequent parts starting from the demarcation point part is changed after the swap, which can avoid the situation where parts of the same size are stacked along the y direction and the layout cannot be effectively changed, and also helps to improve the search efficiency.

[0032] (4) This algorithm has good scalability and can be combined with any population search nesting optimization algorithm. On the basis of constructing the population solution, each solution in the population solution is extended to the dynamic neighborhood involved in this method to improve the search depth of the population search algorithm. Description of the Drawings

[0033] Figure 1 is the overall flowchart of this method;

[0034] Figure 2 are the set conditions for part nesting;

[0035] Figure 3 are the schematic diagrams of the boundary line and the gap in the nesting process;

[0036] Figure 4 In (a), (b), (c), and (d) in are the schematic diagrams of the shape matching degrees f = 0, f = 1, f = 2, and f = 3 generated when the part is placed in the selected gap;

[0037] Figure 5 is the flowchart of the sub-steps for constructing the dynamic neighborhood;

[0038] Figure 6 are the schematic diagrams of the original order, the new order, and their corresponding original layout and new layout when using the traditional algorithm;

[0039] Figure 7 is the comparison schematic diagram of the order adjustment and the layout adjustment when using the traditional algorithm;

[0040] Figure 8 are the schematic diagrams of the front part and the rear part of part p;

[0041] Figure 9 In (a) and (b) in are the schematic diagrams of part p and part q before and after the swap operation in the non-ideal case, and (c) and (d) in are the schematic diagrams of part q after the swap operation in two ideal cases;

[0042] Figure 10 The layout length convergence curve diagram obtained by using the genetic algorithm and the algorithm proposed by the present invention in the said example;

[0043] Figure 11 The layout results obtained by using the genetic algorithm and the algorithm proposed by the present invention in the said example. Detailed implementation manners

[0044] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0045] A layout optimization method based on the best matching rule and tabu search provided by the present invention has an overall flow chart as Figure 1 shown, and the overall process includes the following steps:

[0046] Step 1. Randomly initialize the sequence sequence S and the orientation sequence O, S = {s1, s2, s3,... s n}, O = {o1, o2, o3,... o n}, i ∈ [1, n], where n is the number of rectangular parts. Among them, a random number sequence from 1 to n is initialized as the sequence sequence S of the rectangular parts, and a random number sequence of n 1s or 2s is initialized as the orientation sequence O of the n parts.

[0047] Step 2. Perform layout and create a solution to be adjusted X, that is, input the sequence sequence S and the orientation sequence O of the parts obtained in Step 1 into the rectangular layout algorithm, perform layout, and obtain the actual sequence sequence S' and the actual orientation sequence O', S' = {s'1, s'2, s'3,... s' n}, O' = {o'1, o'2, o'3,... o' n}, and at the same time output the shape matching degree value f corresponding to each part in the layout process this time, and form a shape matching degree sequence F = {f1, f2, f3,... f n}. Define the solution to be adjusted X as a set of the actual sequence sequence, the actual orientation sequence, the shape matching degree sequence, and the occupied length L' of the layout obtained in this layout, that is, X = [S', O', F, L'], and X contains the information of the layout process this time.

[0048] The definition of the rectangular layout is as Figure 2As shown, the width of the sheet is W and the length is L, where W is a fixed value and the length of L is unlimited. The parts are arranged in sequence on the sheet. After the arrangement is completed, the maximum length occupied by the parts in the x direction is L'. The smaller the value of L', the closer the layout of the parts and the higher the nesting quality.

[0049] During the operation of the rectangular nesting algorithm, a series of boundary lines will be generated, forming a series of gaps. The schematic diagrams of the boundary lines and the gaps are as Figure 3 shown. On the left side of the boundary line are the nested parts, and the new part needs to be arranged on the right side of the boundary line. It is set that the gap at the bottom left position is always selected as the position for placing the part, and the part is positioned with its bottom left edge in contact. That is, the left side of the part must be collinear with the left side of the selected gap, and the bottom side of the part must be collinear with the bottom side of the selected gap.

[0050] The rectangular nesting algorithm is a nesting algorithm based on the best matching rule. After each selection of the placement gap for a part, all the remaining parts are traversed to find the part that best matches the shape of the selected gap as the next nested part. Among them, the judgment criterion for the shape matching with the selected gap is whether the part can completely fit the boundary of the selected gap. The selected gap contains three boundaries. When the part is placed in the selected gap, if there is no boundary fit between the part and the selected gap, the shape matching degree is 0. For example Figure 4 as shown in the example of (a) below, if the part has a boundary fit with the selected gap for a certain boundary, the shape matching degree is 1. For example Figure 4 as shown in the example of (b) below, if the part has a boundary fit with the selected gap for two certain boundaries, the shape matching degree is 2. For example Figure 4 as shown in the example of (c) below, if the part has a boundary fit with all three boundaries of the selected gap, the shape matching degree is 3. For example Figure 4 as shown in the example of (d) below. Therefore, the numerical range of the shape matching degree corresponding to all parts is 0 to 3.

[0051] In the nesting algorithm based on the best matching rule, the part that best matches the shape of the selected gap is selected, that is, the part with a shape matching degree of 3 is preferentially selected; if there is no part with a shape matching degree of 3, the part with a shape matching degree of 2 is selected; if there is no part with a shape matching degree of 2 at the same time, the part with a shape matching degree of 1 is selected; if there is no part with a shape matching degree of 1 at the same time, the part with a shape matching degree of 0 is selected. If there are multiple parts with the maximum shape matching degree at the same time, the part that is the most forward in the sequence S is selected.

[0052] During the execution of this nesting process, whenever a certain part is selected and the placement of this part is completed, the actual serial number s' of this part i is stored, the actual orientation o' of this part i and the shape matching degree value f of this part i, then after the arrangement process is completed, the actual sequence S'={s'1,s'2,s'3,…s' n}, the actual position sequence O' = {o'1, o'2, o'3, ... o' n}, shape matching degree sequence F = {f1,f2,f3,…f n}.

[0053] Step 3. Initialize the taboo table T and set the length of the taboo table to n T , and store the solution X. The number of rounds that each solution in the taboo table can be saved during the iterative calculation process is n T .

[0054] Step 4. Select the serial number p of the part to be adjusted in X in combination with the shape matching degree sequence F of X, that is, randomly select an element from all the elements of the shape matching degree sequence F of X that are 0, and let its corresponding part serial number be p. Then the part with serial number p in X is the part to be adjusted. At the same time, part p is used as the dividing point between retention and change in the layout.

[0055] Among them, if the shape matching degree is 1 to 3, it means that the selection of the part is affected by the shape of the selected gap, that is, the reason why the part is selected is that a certain edge (or several edges) of the part can fit with a certain boundary (or several boundaries) of the selected gap, and this selection is conducive to improving the quality of nesting. If the shape matching degree is 0, it means that the selection of the part is not affected by the shape of the selected gap, that is, the reason why the part is selected is only because the part is in the front order, and there is no boundary fit between the part and the selected gap. Therefore, compared with parts with shape matching degrees of 1 to 3, parts with shape matching degrees of 0 are more suitable as parts to be adjusted for subsequent neighborhood operations.

[0056] Step 5. Construct the dynamic neighborhood N of X d Among them, N d Each element n stored in d-i is the combination of the parts order sequence and the orientation sequence, n d-i =[S d-i ,O d-i ], let N d The length of num d .

[0057] Dynamic Neighborhood N d The construction method is to count the total number of different lengths of the side lengths of part p and all parts after part p, and exchange the parts corresponding to each length of the side length with part p. The generated part sequence and orientation sequence together constitute N d , since the total number of edges of different lengths is affected by the choice of position p in the solution to be adjusted and the type of parts after p, therefore, N dIt changes dynamically in each loop.

[0058] Construct the dynamic neighborhood N d The specific process of Figure 5 is as follows, which includes the following sub-steps:

[0059] step1. Save the actual occupied length of part p in the x direction, denoted as l p .

[0060] step2. Establish a list E to store side lengths of different lengths, and add l p to E.

[0061] step3. Initialize the serial number q = p, and the value range of q is [p, n], indicating from part p to part n.

[0062] step4. Judge whether part q meets the exchange condition when using the original angle, that is, if the actual occupied length of part q in the x direction when using the original angle then go to the next step, otherwise, jump to step6.

[0063] step5. Make a copy of the actual sequence S' and the actual orientation sequence O' of X, and exchange part q and part p in them, where part q uses the original angle, and add l q to E, and store the exchanged sequence and orientation sequence in the neighborhood N d , as an element in N d .

[0064] step6. Judge whether part q meets the exchange condition when rotated by 90 degrees, that is, if the actual occupied length of part q in the x direction when rotated by 90 degrees then go to the next step, otherwise, jump to step8.

[0065] step7. Make a copy of the actual sequence S' and the actual orientation sequence O' of X, and exchange part q and part p in them, where part q uses a 90-degree rotation angle, and add l q ' to E, and store the exchanged sequence and orientation sequence in the neighborhood N d , as an element in N d .

[0066] step8. Execute q = q + 1.

[0067] step9. Judge whether the serial number loop is completed. If completed, the process ends; otherwise, return to step4.

[0068] The adjustment measures taken by traditional methods for the part order are usually complex, including operations such as swapping, inserting, and rotating. Taking the genetic algorithm as an example, the genetic algorithm relies on the crossover and mutation steps to generate a new order from the original order. Taking the nesting process of 7 parts a, b, c, d, e, f, g as an example, as Figure 6 shown, assuming the original order is a - b - c - d - e - f - g, the crossover step of the genetic algorithm performs a swapping operation, swapping parts c and e, and the mutation step performs an insertion operation, inserting part d between f and g. Then, after the crossover and mutation steps of the genetic algorithm, the new order is a - b - e - c - f - d - g. The original order and the new order are respectively input into the rectangular nesting algorithm to perform nesting, and the corresponding nesting layouts can be obtained respectively, as Figure 6 shown by the "original layout" and "new layout" in. Due to the characteristic of backward influence in the nesting process, when the third part in the original order changes from c to e, the arrangement of part e will change the original part boundary line information, thereby affecting the arrangement of subsequent parts, causing the layouts of subsequent parts c, f, d, g to all change. Therefore, for the new layout of the 7 parts compared to the original layout, the layouts of all parts starting from the third part have changed. For the above 7 parts, in terms of order, the parts remaining unchanged are a, b, g, and the parts changed are c, d, e, f. And in terms of layout, the parts remaining unchanged are a, b, and the parts changed are c, d, e, f, g, as Figure 7 shown, the adjustment in order cannot correspond to the adjustment in layout. Thus, it can be seen that the characteristic of backward influence in the nesting process restricts the effectiveness of the adjustment in order. The complex operations taken by traditional methods for the part order cannot produce corresponding layout adjustment effects, and the complex operations taken are unnecessary and inefficient operations. Therefore, traditional methods have the defect of insufficient search efficiency.

[0069] The present invention changes the way of order update, that is, selects a certain part in the original order as the demarcation point for retention and change in layout, and avoids the complex operations in order adjustment, but only takes the swapping operation between this part and its subsequent parts. Taking Figure 8Taking the situation shown as an example, all the pre-components of component p are on the left side of the component boundary line where the placement gap selected for component p is located, while on the right side of the boundary line are component p and all its post-components. Taking component p as the demarcation point for retention and change in the layout, swapping component p with its post-component q, the layout of the pre-components of component p remains unchanged while the layout of component p and its post-components is adjusted. In this way, the layout of the components before the demarcation point component is retained, and the layout of the components after the demarcation point component is changed. This method is more in line with the way of component layout update (that is, the characteristic of component layout update is to use a certain component as the segmentation point for retention and change, the layout of the components before is retained, and the layout of the components after is changed), and it avoids the excessive unnecessary and inefficient complex operations adopted by the traditional method. Therefore, the search efficiency is higher than that of the traditional method.

[0070] Meanwhile, select the component with a shape matching degree value of 0 as the demarcation point, that is, only select the components that cannot match the shape of the boundary line as the adjustment objects, which can retain the layout of the components with higher shape matching degree values and fully retain the positioning efficiency of the best matching rule.

[0071] In addition, select the component with a different size in the x direction from the demarcation point component as the swapping object, which can ensure that the layout of the subsequent components starting from the demarcation point component is changed after swapping. Taking Figure 9 the situation shown as an example, if the size condition is not set when selecting the component as the swapping object, as shown by component p and component q in (a) of Figure 9 , after swapping, component p and component q are still arranged in a stacked manner along the y direction, as shown by (b) of Figure 9 , then the layout of the subsequent components starting from the demarcation point component does not change effectively. And setting the condition that the components have different sizes in the x direction can avoid the non-ideal situations shown by (a) and (b) of Figure 9 , ensure that the layout of the subsequent components starting from the demarcation point component is changed after swapping, as shown by (c) and (d) of Figure 9 . Therefore, this set condition also helps to improve the search efficiency.

[0072] Step 6. Execute nesting and create a dynamic neighborhood solution set Z d , that is, input each element n d in N d-i into the rectangular nesting algorithm and execute the nesting process to obtain the actual component sequence S’ d-i , actual orientation sequence O’ d-i , shape matching degree sequence F d-i , and length L’ d-i combination corresponding to this element, and store the combination of these four pieces of information into Z d as the i-th element in it, that is, z d-i = [S’d-i , O’ d-i , F d-i , L’ d-i .

[0073] After the creation of the dynamic neighborhood solution set Z d ends, the length of Z d is the same as the length of N d , both are num d , and the process of performing nesting is the same as in Step 2.

[0074] Step 7. Update the solution X to be adjusted, that is, select the solution in the neighborhood solution set Z d with the shortest occupied length and not in the taboo list T as the new solution X to be adjusted.

[0075] Step 8. Update the taboo list T. If the number of solutions already stored in T is less than n T , then store the solution X to be adjusted at the end of T. If the number of solutions already stored in T is greater than or equal to n T , then remove the solution at the beginning of the table, move all the stored solutions one unit towards the beginning of the table, and store the solution X to be adjusted at the end of T.

[0076] Step 9. Determine whether the set number of loops n c is reached. If the number of loops reaches n c , the process ends. Otherwise, return to Step 4.

[0077] In view of the problem that the traditional search algorithm has insufficient search efficiency due to the characteristic of not considering the backward influence in the nesting process, the present invention provides a nesting optimization method based on the best matching rule and taboo search. A new method of selecting a certain part as the demarcation point for retention and change in the layout in the original order is adopted, and the high-efficiency positioning advantage of the best matching rule can be fully utilized. At the same time, the situation where parts of the same size are stacked along the y direction and the layout cannot produce effective changes can be avoided, and the efficiency of nesting optimization is improved.

[0078] The following uses a group of examples to verify the effectiveness of the algorithm proposed by the present invention. It is set that the width W of the sheet is 100, the value range of the part specifications is set, the length range of the parts is [10, 30], and the width range is [5, 15]. 50 rectangular parts with lengths and widths within the given range are randomly generated. The shortest nesting lengths of this group of parts are calculated respectively using the genetic algorithm and the algorithm proposed by the present invention. Each of the two algorithms calculates 2000 solutions. During the execution of the algorithms, the process of the shortest nesting length obtained changing with the number of solutions searched is recorded, and the average value of 100 execution processes is obtained. The nesting length convergence curves of the two algorithms can be obtained, as shown in Figure 10 . In a certain execution process, the nesting results obtained by the genetic algorithm and the algorithm proposed by the present invention are as shown in Figure 11As shown. By Figure 10 and Figure 11 From the comparison results in, it can be seen that the search efficiency of the algorithm proposed by the present invention is better than that of the genetic algorithm.

[0079] It is easy for those skilled in the art to understand that the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A layout optimization method based on the best matching rule and tabu search, characterized in that Including: S1. Randomly initialize the sequence and orientation sequence of rectangular parts; S2. Perform a rectangular nesting algorithm based on the best matching rule on the initialized rectangular parts to obtain the solution of the nesting information X ; The solution of the nesting information includes the actual order sequence, the actual orientation sequence, the shape matching degree sequence composed of the shape matching degree values corresponding to each part, and the layout occupation length obtained from this nesting S3. Solve from the nesting information X Select any part with a shape matching value of 0 as the part to be adjusted p ; S4. Move the part located after the part to be adjusted p and different from p in x the direction with the part having different dimensions, and p perform an exchange operation, regenerate the actual part sequence, actual orientation sequence, shape matching degree sequence, and layout occupied length, and construct a dynamic neighborhood solution set; S5. Select the solution with the shortest layout occupation length and different from the nesting information solution corresponding to the historical iteration round from the dynamic neighborhood solution set as the new nesting information solution; S6. Determine whether the set number of loops is reached; if not, return to execute step S3; if so, take the layout corresponding to the current nesting information solution as the final nesting result; The nesting algorithm based on the best matching rule specifically includes the following process: Select the gap at the bottom leftmost position as the position to place the part, and use the bottom left abutting positioning; Traverse all the remaining parts, and find the part that best matches the shape of the selected gap as the next nesting part; If there are multiple parts with the maximum shape matching degree at the same time, select the part that is the foremost in the sequence; The judgment criterion for matching the shape of the selected gap is whether the part can completely fit the boundary of the selected gap; if the part does not have a boundary fit with the selected gap, the shape matching degree is 0; if the part has a boundary fit with the selected gap, the shape matching degree is 1; if the part has two boundary fits with the selected gap, the shape matching degree is 2; if the part has all three boundaries fitting with the selected gap, the shape matching degree is 3.

2. The layout optimization method based on the best matching rule and tabu search according to claim 1, wherein, The specific process of performing the swap operation in step S4 is: (1) Determine the position of the part to be adjusted p After the parts q When placed at the original angle x Direction actual length and part p exist x Are the actual lengths occupied by the directions the same? If they are the same, perform the exchange operation; otherwise, go to (2); (2)Judge the part q When rotating 90 degrees, check if the actual occupied length in the x direction is the same as the actual occupied length of the part in the p in the x direction; if they are the same, perform the swapping operation; if not, sequentially select the next part and go to (1).

3. A layout optimization system based on the best matching rule and tabu search, characterized in that, Including: Initialization module, randomly initialize the sequence and orientation sequence of rectangular parts; Initial nesting module, which performs a rectangular nesting algorithm based on the best matching rule on the initialized rectangular parts to obtain the solution of nesting information X ; The solution of the nesting information includes an actual order sequence, an actual orientation sequence, a shape matching degree sequence composed of the shape matching degree values corresponding to each part, and the layout occupation length obtained from this nesting Shape matching degree screening module, from the nesting information solution X Select any part with a shape matching degree value of 0 as the part to be adjusted p ; The dynamic neighborhood solution set construction module places the parts to be adjusted p after and p in the x direction with different dimensions of the parts and p perform an exchange operation to regenerate the actual order sequence, actual orientation sequence, shape matching degree sequence, and layout occupied length of the parts, and construct a dynamic neighborhood solution set; Tabu search module, select the solution with the shortest layout occupation length and different from the nesting information solution corresponding to the historical iteration round from the dynamic neighborhood solution set as the new nesting information solution; Loop judgment module, determine whether the set number of loops is reached; if not, return to the shape matching degree screening module; if so, take the layout corresponding to the current nesting information solution as the final nesting result; The nesting algorithm based on the best matching rule specifically includes the following process: Select the gap at the bottom leftmost position as the position to place the part, and use the bottom left abutting positioning; Traverse all the remaining parts, and find the part that best matches the shape of the selected gap as the next nesting part; If there are multiple parts with the maximum shape matching degree at the same time, select the part that is the foremost in the sequence; The judgment criterion for matching the shape of the selected gap is whether the part can completely fit the boundary of the selected gap; if the part does not have a boundary fit with the selected gap, the shape matching degree is 0; if the part has a boundary fit with the selected gap, the shape matching degree is 1; if the part has two boundary fits with the selected gap, the shape matching degree is 2; if the part has all three boundaries fitting with the selected gap, the shape matching degree is 3.

4. The layout optimization system based on the best matching rule and tabu search according to claim 3, characterized in that The specific process of the dynamic neighborhood solution set construction module performing the swap operation is: (1) Determine the parts located after the part p to be adjusted q When placed at the original angle, x Whether the actual occupied length in the p direction is the same as the actual occupied length of the part x in the direction; if the same, perform the exchange operation; otherwise, go to (2); (2)Judge the part q When rotating 90 degrees, at x Whether the actual occupied length in the direction is the same as that of the part p At x Whether the actual occupied length in the direction is the same; if it is the same, perform the exchange operation; if not, sequentially select the next part and go to (1).

5. A computer-readable medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method according to claim 1 or 2 are implemented.

Images