Mesh process route-oriented similarity measurement method and application thereof in CAPP
By converting the mesh process route into a set of linear process routes, and using multi-level pseudo LCS similarity, Shannon entropy similarity, Jaccard similarity and structural complexity similarity to quantify process information, combined with the KM algorithm and PCA method, the problem of inaccurate similarity measurement of mesh process routes is solved, and the retrieval and planning efficiency of the CAPP system is improved.
Patent Information
- Application Number
- CN202510745112.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-05
- Publication Date
- 2025-09-19
AI Technical Summary
In the existing technology, the structure of the mesh process route is highly complex, resulting in inaccurate similarity measurement and the inability to effectively utilize the process information in its complex structure.
The mesh process route is converted into a linear process route set. The local element information is quantified by multi-level pseudo LCS similarity, Shannon entropy similarity, Jaccard similarity and structural complexity similarity. The KM algorithm and PCA method are combined to form a comprehensive similarity index to comprehensively measure the process information.
It achieves fine granularity and sensitivity of the mesh process route, improves the accuracy and efficiency of the CAPP system in retrieving and planning the mesh process route, and reduces the difficulty of modification and planning for technical personnel.
Smart Images
Figure CN120670864A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of intelligent manufacturing, and in particular relates to a similarity measurement method for a mesh process route and its application in CAPP. Background Art
[0002] In the manufacturing sector, process planning refers to the process of designing a rational process route for machining workpieces, taking into account specific production conditions and ensuring that it meets various production requirements (such as process requirements, economic requirements, and environmental requirements). In recent years, as manufacturing has entered an era of intelligent and knowledge-based processes, a large amount of manufacturing data has been stored and wasted. However, this data actually represents explicit knowledge, reflecting the experience and preferences of technicians. Therefore, effectively discovering and reusing this knowledge is a key factor in improving corporate competitiveness.
[0003] Computer-aided process planning (CAPP) systems are key tools for automating the generation and management of process routes in the manufacturing industry. One of the core tasks of CAPP systems is to compare and optimize different process routes, and similarity measurement is fundamental to this task. By calculating the similarity between process routes, it is possible to reuse, optimize, and recommend them, thereby improving production efficiency and resource utilization.
[0004] In the current manufacturing field, the study of process route similarity is of great significance, especially in terms of improving production efficiency, reducing costs, and optimizing resource allocation. At present, experts and scholars have achieved relatively fruitful results in the study of linear process route similarity. For example, Liu et al. established a process route similarity measurement relationship based on Euclidean distance, and Zhang Hui et al. constructed a multi-level similarity calculation model for processing process routes. These studies mainly focus on how to use quantitative methods to evaluate the similarity between linear process routes and, on this basis, achieve process route optimization and reuse. However, compared with linear process routes, there are relatively few related studies on the similarity design of mesh process routes. Due to its complex structure and diverse path selection, the measurement and optimization of similarity of mesh process routes becomes more difficult. Despite this, mesh process routes are more in line with the current needs of flexible manufacturing. Flexible manufacturing emphasizes the flexibility and adaptability of production and can quickly respond to market changes and customer needs.
[0005] Because mesh process routes have a more complex structure and require more process information to be considered, the similarity metrics used in existing linear process route designs cannot be directly applied to mesh process routes. Currently, some existing technologies aim to extract and push typical process routes for discrete manufacturing systems, and conduct research on the clustering problem of sequential process routes and mesh process routes in discrete manufacturing systems. On the basis of unified coding of process routes, they consider the same process information and the ordering information of the processes, and propose a new comprehensive indicator to describe the similarity between process routes, thereby constructing a dissimilarity matrix. Navaei and Elmaraghy were the first to study mesh process routes. In order to classify variants of parts or products, they designed a new mesh process route similarity coefficient, which comprehensively considers the sequential relationship, the number of the same processes, and the volume similarity, and combines them through a linear weighted sum.
[0006] Currently, similarity metrics for mesh process routes do not consider the similarity between processes. Previous studies typically default the similarity between all processes to zero. In reality, there is some similarity between different processes, which influences the overall similarity between mesh process routes. Furthermore, the structural complexity of mesh process routes is a key factor, which is absent in linear process routes. Existing technologies lack targeted improvements in this regard.
[0007] Due to the high structural complexity of mesh process routes, the accuracy of similarity metrics for process routes with significant structural differences requires a more detailed analysis. The similarity metric setting simply considers the number of individual process routes within a mesh process route, without taking into account the length of each individual route. This is also a problem that needs to be addressed. Summary of the Invention
[0008] The purpose of the present invention is to provide a similarity measurement method for mesh process routes and its application in CAPP, which is used to solve the technical problem in the prior art that the similarity measurement of process routes is not accurate enough due to the high structural complexity of mesh process routes and the large differences in structural complexity.
[0009] The similarity measurement method for a mesh process route includes the following steps:
[0010] Step 1: Process information analysis and quantification: The meshed process route is converted into a set of linear process routes. Based on this, the local element information and the overall information of the linear process route set are analyzed. The local element information refers to the common information between the linear process routes in the set, including the similarity between processes, the sequence relationship, and the repetition frequency of processes. The overall information of the set includes the number of identical process routes and the structural complexity.
[0011] Step 2: Integration and fusion of intermediate similarities: The above four different intermediate similarity metrics are combined into a composite indicator through the KM algorithm and PCA method.
[0012] Preferably, step 1 specifically includes:
[0013] Step 1: Analyze the network process route from the perspective of linear process route set, and design intermediate similarity. The intermediate similarity metrics include multi-level pseudo LCS similarity, Shannon entropy similarity, Jaccard similarity, and structural complexity similarity, which are used to measure the sequence relationship, process repetition frequency, number of identical process routes, and structural complexity, respectively.
[0014] Step 2: Calculate the similarity between processes to measure the similarity between processes;
[0015] Step 3: Quantify the order relationship between linear process routes based on multi-level pseudo LCS similarity;
[0016] Step 4: Use Shannon entropy similarity to quantify the process repetition frequency;
[0017] Step 5: Use Jaccard similarity to quantify the number of identical process routes;
[0018] Step 6: Quantify the structural complexity using structural complexity similarity based on the number of optional process routes.
[0019] Preferably, Step 2 includes: using a three-level code to represent each process, and based on this, calculating the similarity between processes through an XOR operation, the formula is as follows:
[0020]
[0021] Among them i and o j There are two processes, and are the first k bits of the codes for these two processes respectively.
[0022] Preferably, Step 3 includes:
[0023] Step 3.1: Extract the pseudo LCS between two linear process routes based on dynamic programming and similarity between processes. Each extracted pseudo LCS is denoted as L i , i represents the number of extractions
[0024] Step 3.2: Remove L from the two linear process routes M and N i , the remaining processes are defined as two new linear process routes M and N;
[0025] Step 3.3: If L i If it is not empty or the lengths of M and N are not 0, return to Step 3.1 and repeat the calculation to obtain a new first-level pseudo LCS. Otherwise, go to Step 3.4.
[0026] Step 3.4: Calculate the corresponding multi-level pseudo LCS similarity S based on the previously generated multi-level pseudo LCS L , the calculation formula is as follows:
[0027]
[0028] Where q represents the maximum number of stages, u and v are the lengths of the original process routes M and N, respectively, and L is the LCS of the two linear process routes M and N.
[0029] Preferably, Step 3.1 includes:
[0030] ① Calculate the dynamic programming matrix between two linear process routes M and N. The calculation formula is as follows:
[0031]
[0032] Among them, j, k are the jth and kth processes,
[0033] ② Find the pseudo LCS from the dynamic programming matrix through the backtracking method, denoted as L i ;
[0034] ③ Based on pseudo-LCSL i Calculate PR i ,pr i is the sum of the process similarities of the pseudo LCS, and the corresponding algorithm is as follows:
[0035] pr i =∑s o ,
[0036] s o It represents the similarity between any two processes.
[0037] Preferably, Step 4 includes:
[0038] Step 4.1: The xth process in LCS The density of the jth occurrence in process route M is expressed as The expression is as follows:
[0039]
[0040] in, It is the position where the process appears for the jth time, and the initial value p0 is defined as 0. rpt refers to the number of repetitions of the process in the corresponding linear process route.
[0041] Step 4.2: Describe the xth process in LCS The density of the i-th occurrence is calculated as follows:
[0042]
[0043] in, Indicates the xth process in LCS The density of the kth occurrence in process route M;
[0044] Step 4.3: Calculate the Shannon entropy of each process in the LCS at each position in the corresponding linear process route The calculation formula is as follows:
[0045]
[0046] Step 4.4: Calculate the Shannon entropy similarity: Calculate the similarity between process routes M and N based on the Shannon entropy. The formula is as follows:
[0047]
[0048] in Indicates the process in process route M Calculation results of Shannon entropy (process in LCS); Indicates process N in process route The calculation results of Shannon entropy;
[0049] Step 4.5: A correction factor that reflects inconsistent lengths is introduced to establish the final Shannon entropy similarity calculation formula as follows:
[0050]
[0051] max(u,v) is a correction factor that reflects inconsistent lengths.
[0052] Preferably, in Step 5, the Jaccard similarity is calculated as follows:
[0053]
[0054] E and F are sets of linear process routes transformed from different network process routes.
[0055] Preferably, in Step 6, the structural complexity similarity S S The calculation formula is as follows:
[0056]
[0057] Among them, op E and op Fis the number of optional process routes for mesh process routes E and F.
[0058] Preferably, step 2 specifically includes:
[0059] Step 1: Convert the similarity between linear process route elements into the similarity between linear process route sets;
[0060] Step 2: Considering the correlation between the four intermediate similarities, the PCA method is used to fuse the four intermediate similarities to obtain the comprehensive similarity.
[0061] The present invention also provides an application of the above-mentioned similarity measurement method for network process routes in CAPP, comprising the following steps:
[0062] Step 1: Obtain current product design requirements;
[0063] Step 2: Calculate the comprehensive similarity of the historical mesh process routes using the aforementioned similarity measurement method for mesh process routes; search the historical mesh process route database for the most similar historical mesh process route based on the comprehensive similarity, and then output it to the technician;
[0064] Step 3: The technicians adjust and modify the historical mesh process route according to the process constraints and actual needs to obtain a new mesh process route that meets the needs, and store the new mesh process route in the database for future retrieval;
[0065] Step 4: Put the new mesh process route into production.
[0066] The present invention has the following advantages: the present invention analyzes the mesh process route and converts it into a set of linear process routes. In the similarity measurement, the present invention comprehensively considers various process information for the more complex structure of the mesh process route, and analyzes the process information from both local and overall aspects, including the similarity between processes, the sequence relationship, the process repetition frequency, the number of identical process routes and the structural complexity. For the sequence relationship, the process repetition frequency, the number of identical process routes and the structural complexity that are difficult to measure directly, the present method sets four intermediate similarities for measurement, thereby quantifying the above-mentioned various process information through reasonable similarity design.
[0067] On the basis of realizing the above-mentioned industrial information measurement, the present invention also provides a set-oriented similarity measurement framework to integrate the above-mentioned quantitative indicators into a comprehensive similarity measurement, so as to more effectively realize a more fine-grained and sensitive mesh process route similarity measurement for the corresponding mesh process route (corresponding to the linear process route set).
[0068] Based on the above-mentioned similarity measurement method, the present invention applies it to CAPP, effectively improving the accuracy and quality of the historical process routes retrieved by CAPP, thereby greatly reducing the difficulty and time for relevant technical personnel to perform process modification and planning. BRIEF DESCRIPTION OF THE DRAWINGS
[0069] Figure 1 The present invention is a flowchart of a similarity measurement method for a mesh process route and its application in CAPP.
[0070] Figure 2 Schematic diagram of the process of step 1 in the present invention.
[0071] Figure 3 Schematic diagram of the process of step 2 in the present invention. DETAILED DESCRIPTION
[0072] The specific implementation methods of the present invention will be further explained in detail below through the description of embodiments with reference to the accompanying drawings, so as to help those skilled in the art to have a more complete, accurate and in-depth understanding of the inventive concept and technical solution of the present invention.
[0073] like Figure 1-Figure 3 As shown, the present invention provides a similarity measurement method for a mesh process route, comprising the following steps.
[0074] Step 1: Process information analysis and quantification: The mesh process route is converted into a linear process route set, and based on this, the local element information and the overall information of the set in the linear process route set are analyzed.
[0075] In this solution, the mesh process route is essentially considered as a collection of several linear process routes. Therefore, to consider the process information contained in the mesh process route from the perspective of the linear process route collection, that is, the similarity between the two collections, it is necessary to consider it from both the overall and local levels. Figure 2 As shown in the figure, the so-called local level refers to the common information between linear process routes in the set, including inter-process similarity, sequence relationships, and process repetition frequency. From the perspective of the entire set, the main considerations are the number of identical process routes and structural complexity. Based on the local element information, inter-process similarity, sequence relationships, and process repetition frequency are quantified; based on the overall information of the set, the number of identical process routes and structural complexity are quantified. Finally, four different intermediate similarity metrics are obtained: multi-level pseudo-LCS similarity, Shannon entropy similarity, Jaccard similarity, and the number of optional process routes.
[0076] Based on the analysis results, four quantitative measurement methods for information are proposed. The specific steps for analyzing process information are as follows:
[0077] Step 1: Analyze the mesh process route from the perspective of linear process route set, and design the intermediate similarity for reasonable measurement and quantification (see Figure 1 ).
[0078] During the analysis process, the mesh process route is converted into a set of linear process routes to obtain the process information of each linear process route. The process information includes the similarity between processes, the sequence relationship, the process repetition frequency, the number of identical process routes and the structural complexity. The intermediate similarities include multi-level pseudo LCS similarity, Shannon entropy similarity, Jaccard similarity and structural complexity similarity, which are used to measure the sequence relationship, the process repetition frequency, the number of identical process routes and the structural complexity, respectively.
[0079] Step 2: Calculate the similarity between processes to measure the similarity between processes.
[0080] Based on the national standard JB / T5992.2-5992.10, a three-level code is used to represent each process (such as annealing is 511, turning is 311, etc.), and the similarity between processes is calculated based on this through the exclusive-or operation (XNOR). The formula is as follows:
[0081]
[0082] Among them i and o j There are two processes, and are the first k bits of the codes for these two processes respectively.
[0083] Step 3: Based on multi-level pseudo LCS similarity S L To quantify the order relationship between linear process routes. The specific steps are as follows.
[0084] Step 3.1: Extract the pseudo LCS (longest subsequence string) between two linear process routes based on dynamic programming and process similarity. During the initial extraction, i = 1, where i represents the number of extractions. This step also includes:
[0085] ① Calculate the dynamic programming matrix between two linear process routes M and N. The calculation formula is as follows:
[0086]
[0087] Among them, j, k are the jth and kth processes,
[0088] ② Find the pseudo LCS from the dynamic programming matrix through the backtracking method, denoted as L i .
[0089] ③ Based on pseudo-LCSL i Calculate PRi ,pr i is the sum of the process similarities of the pseudo-LCS, namely pseudo-r i The corresponding algorithm is as follows:
[0090] pr i =∑s o ,
[0091] s o It represents the similarity between any two processes.
[0092] Step 3.2: Remove L from the two linear process routes M and N i , the remaining processes are defined as two new linear process routes M and N.
[0093] Step 3.3: If L i If it is not empty or the lengths of M and N are not 0, return to Step 3.1 and repeat the calculation to obtain a new level of pseudo LCS. At this time, the number of times the pseudo LCS is extracted increases by 1, that is, i = i + 1. Otherwise, go to Step 3.4.
[0094] Step 3.4: Calculate the corresponding multi-level pseudo LCS similarity S based on the previously generated multi-level pseudo LCS L , the calculation formula is as follows:
[0095]
[0096] Where q represents the maximum number of stages, u and v are the lengths of the original process routes M and N, respectively, and L is the LCS of the two linear process routes M and N.
[0097] Step 4: Use Shannon entropy similarity S E To quantify the process repetition frequency, the specific steps are as follows:
[0098] Step 4.1: Represent the repeated processes in the linear process route: the xth process in LCS The density of the jth occurrence in process route M is expressed as The expression is as follows:
[0099]
[0100] in, It is the position where the process appears for the jth time. The initial value p0 is defined as 0. rpt refers to the number of repetitions of the process in the corresponding linear process route.
[0101] Step 4.2: Describe the xth process in LCS The density of the i-th occurrence is calculated as follows:
[0102]
[0103] in, Indicates the xth process in LCS The density of the kth occurrence in process route M.
[0104] Step 4.3: Calculate the Shannon entropy of each process in the LCS at each position in the corresponding linear process route The calculation formula is as follows:
[0105]
[0106] Step 4.4: Calculate the Shannon entropy similarity: Calculate the similarity between process routes M and N based on the Shannon entropy. The formula is as follows:
[0107]
[0108] in Indicates the process in process route M Calculation results of Shannon entropy (process in LCS); Indicates process N in process route The calculation results of Shannon entropy.
[0109] Step 4.5: A correction factor that reflects inconsistent lengths is introduced to establish the final Shannon entropy similarity calculation formula as follows:
[0110]
[0111] Considering that the lengths of two linear process routes may be different in practice, this step introduces a correction factor max(u,v) that can reflect the inconsistent lengths.
[0112] Step 5: Use Jaccard similarity S J Quantify the number of identical process routes. The Jaccard similarity is calculated as follows:
[0113]
[0114] E and F are sets of linear process routes transformed from different network process routes.
[0115] Step 6: Use the structural complexity similarity S based on the number of optional process routes S Quantify the structural complexity. Structural complexity similarity S S The calculation formula is as follows:
[0116]
[0117] Among them, op E and opF is the number of optional process routes for the mesh process routes E and F. Here, the number of optional process routes in the mesh process route is used to quantify the structural complexity, and the structural complexity similarity of the two mesh process routes is calculated based on this, thereby quantifying the similarity between the process routes in terms of structural complexity.
[0118] Step 2: Integration and fusion of intermediate similarities: The above four different intermediate similarity metrics are combined into a composite indicator through the KM algorithm (Kuhn-Munkres algorithm) and PCA (principal component analysis) method.
[0119] The four intermediate similarity metrics obtained above calculate the process information from different aspects, but these four similarity metrics cannot be applied directly and need to be integrated into a single comprehensive similarity through reasonable means. In order to obtain this comprehensive similarity index, this patent designs a set-oriented similarity measurement integration framework to combine the above four different intermediate similarity metrics into a composite index to ensure the accuracy and effectiveness of the similarity measurement for the mesh process route. The overall design framework is as follows: Figure 3 The specific steps for synthesizing composite indicators are as follows:
[0120] Step 1: Convert the similarity between linear process route elements into the similarity between linear process route sets. The specific steps are as follows.
[0121] Step 1.1: Based on the multi-level pseudo LCS similarity S obtained in step 1 L and Shannon entropy similarity S E , construct the similarity matrix between elements.
[0122] Step 1.2: Calculate the similarity between each similarity matrix based on the KM algorithm, and take the maximum weight as the similarity between the two similarity matrices;
[0123] Step 1.3: Divide the similarity between the similarity matrices by the value of the maximum number of elements in the two linear process route sets to obtain the similarity value between the two linear process route sets.
[0124] Step 2: Considering the correlation between the four intermediate similarities, the PCA method is used to fuse the four intermediate similarities to obtain the comprehensive similarity as a more accurate comprehensive indicator.
[0125] The present invention also provides an application of the above-mentioned similarity measurement method for a network process route in CAPP. Based on the above steps 1 and 2, the application further includes:
[0126] The present invention also provides an application of the aforementioned similarity measurement method in CAPP, providing a practical application framework for applying the proposed comprehensive similarity measurement of networked process routes to the CAPP retrieval process, thereby improving the accuracy and quality of CAPP in retrieving networked process routes. Specifically, the framework includes the following steps.
[0127] Step 1: Obtain current product design requirements.
[0128] Step 2: Calculate the comprehensive similarity of the historical mesh process routes using the aforementioned similarity measurement method for mesh process routes; search for the most similar historical mesh process routes from the historical mesh process route database based on the comprehensive similarity, and then output the result to the technical staff.
[0129] Step 3: The technicians adjust and modify the historical mesh process route according to the process constraints and actual needs to obtain a new mesh process route that meets the needs, and store the new mesh process route in the database for the next retrieval.
[0130] Step 4: Put the new mesh process route into production.
[0131] The present invention is described above by way of example in conjunction with the accompanying drawings. It is obvious that the specific implementation of the present invention is not limited to the above-mentioned method. As long as various non-substantial improvements are made using the inventive concept and technical solution of the present invention, or the inventive concept and technical solution are directly applied to other occasions without improvement, they are all within the scope of protection of the present invention.
Claims
1. A similarity measurement method for a mesh process route, characterized by: The following steps are involved: Step 1: Process information analysis and quantification: The meshed process route is converted into a set of linear process routes. Based on this, the local element information and the overall information of the linear process route set are analyzed. The local element information refers to the common information between the linear process routes in the set, including the similarity between processes, the sequence relationship, and the repetition frequency of processes. The overall information of the set includes the number of identical process routes and the structural complexity. Step 2: Integration and fusion of intermediate similarities: The above four different intermediate similarity metrics are combined into a composite indicator through the KM algorithm and PCA method.
2. The similarity measurement method for a mesh process route according to claim 1, characterized in that: Step 1 specifically includes: Step 1: Analyze the network process route from the perspective of linear process route set, and design intermediate similarity. The intermediate similarity metrics include multi-level pseudo LCS similarity, Shannon entropy similarity, Jaccard similarity, and structural complexity similarity, which are used to measure the sequence relationship, process repetition frequency, number of identical process routes, and structural complexity, respectively. Step 2: Calculate the similarity between processes to measure the similarity between processes; Step 3: Quantify the order relationship between linear process routes based on multi-level pseudo LCS similarity; Step 4: Use Shannon entropy similarity to quantify the process repetition frequency; Step 5: Use Jaccard similarity to quantify the number of identical process routes; Step 6: Quantify the structural complexity using structural complexity similarity based on the number of optional process routes.
3. The similarity measurement method for a network process route according to claim 2, characterized in that: Step 2 includes: using three-level codes to represent each process, and based on this, calculating the similarity between processes through the same or same operation, the formula is as follows: Among them i and o j There are two processes, and are the first k bits of the codes for these two processes respectively.
4. The similarity measurement method for a network process route according to claim 3, characterized in that: Step 3 includes: Step 3.1: Extract the pseudo LCS between two linear process routes based on dynamic programming and similarity between processes. Each extracted pseudo LCS is denoted as L i , i represents the number of extractions Step 3.2: Remove L from the two linear process routes M and N i , the remaining processes are defined as two new linear process routes M and N; Step 3.3: If L i If it is not empty or the lengths of M and N are not 0, return to Step 3.1 and repeat the calculation to obtain a new first-level pseudo LCS. Otherwise, go to Step 3.
4. Step 3.4: Calculate the corresponding multi-level pseudo LCS similarity S based on the previously generated multi-level pseudo LCS L , the calculation formula is as follows: Where q represents the maximum number of stages, u and v are the lengths of the original process routes M and N, respectively, and L is the LCS of the two linear process routes M and N.
5. The similarity measurement method for a network process route according to claim 4, characterized in that: Step 3.1 includes: ① Calculate the dynamic programming matrix between two linear process routes M and N. The calculation formula is as follows: Among them, j, k are the jth and kth processes, ② Find the pseudo LCS from the dynamic programming matrix through the backtracking method, denoted as L i ; ③ Based on pseudo-LCSL i Calculate PR i ,pr i is the sum of the process similarities of the pseudo LCS, and the corresponding algorithm is as follows: pr i =∑s o , s o It represents the similarity between any two processes.
6. The similarity measurement method for a network process route according to claim 2, characterized in that: Step 4 includes: Step 4.1: The xth process in LCS The density of the jth occurrence in process route M is expressed as The expression is as follows: in, It is the position where the process appears for the jth time, and the initial value p0 is defined as 0. rpt refers to the number of repetitions of the process in the corresponding linear process route. Step 4.2: Describe the xth process in LCS The density of the i-th occurrence is calculated as follows: in, Indicates the xth process in LCS The density of the kth occurrence in process route M; Step 4.3: Calculate the Shannon entropy of each process in the LCS at each position in the corresponding linear process route The calculation formula is as follows: Step 4.4: Calculate the Shannon entropy similarity: Calculate the similarity between process routes M and N based on the Shannon entropy. The formula is as follows: in Indicates the process in process route M Calculation results of Shannon entropy (process in LCS); Indicates process N in process route The calculation results of Shannon entropy; Step 4.5: A correction factor that reflects inconsistent lengths is introduced to establish the final Shannon entropy similarity calculation formula as follows: max(u,v) is a correction factor that reflects inconsistent lengths.
7. The similarity measurement method for a network process route according to claim 2, characterized in that: In Step 5, the Jaccard similarity is calculated as follows: E and F are sets of linear process routes transformed from different network process routes.
8. The similarity measurement method for a mesh process route according to claim 2, characterized in that: In Step 6, the structural complexity similarity S S The calculation formula is as follows: Among them, op E and op F is the number of optional process routes for mesh process routes E and F.
9. The similarity measurement method for a mesh process route according to claim 1, characterized in that: Step 2 specifically includes: Step 1: Convert the similarity between linear process route elements into the similarity between linear process route sets; Step 2: Considering the correlation between the four intermediate similarities, the PCA method is used to fuse the four intermediate similarities to obtain the comprehensive similarity.
10. Application of a similarity measurement method for a network process route according to any one of claims 1 to 9 in CAPP, characterized in that: The following steps are involved: Step 1: Obtain current product design requirements; Step 2: Calculating the comprehensive similarity of the historical mesh process routes by using a similarity measurement method for mesh process routes according to any one of claims 1 to 9; Based on the comprehensive similarity, the most similar historical mesh process route is searched from the historical mesh process route database and then output to the technicians; Step 3: The technicians adjust and modify the historical mesh process route according to the process constraints and actual needs to obtain a new mesh process route that meets the needs. The new mesh process route is stored in the database for future retrieval. Step 4: Put the new mesh process route into production.