Intelligent dependency graph construction method based on multi-dimensional analysis and adaptive optimization

Abstract

The invention relates to data analysis, in particular to an intelligent dependency graph construction method based on multi-dimensional analysis and self-adaptive optimization, which comprises the following steps: reading module data, and carrying out preliminary analysis on a dependency relationship among modules to construct a preliminary dependency graph; performing multi-dimensional analysis on a dependency relationship between each pair of modules in the preliminary dependency graph, and calculating a comprehensive score of each pair of modules according to a multi-dimensional analysis result; adjusting the preliminary dependency graph according to the comprehensive score of each pair of modules to obtain an adjusted dependency graph; performing cyclic dependency detection on the adjustment dependency graph, and merging static dependency data and dynamic dependency data of each module; self-adaptive optimization is carried out according to the complexity of the adjusted dependency graph, a final dependency graph is generated, and the final dependency graph is stored and output; according to the technical scheme provided by the invention, the defect that the software module dependency graph is difficult to accurately and efficiently analyze and construct in the prior art can be effectively overcome.

Description

Technical Field

[0001] The present invention relates to data analysis, and in particular to an intelligent dependency graph construction method based on multidimensional analysis and adaptive optimization. Background Art

[0002] As the complexity of software systems continues to increase, it becomes increasingly important to understand and maintain the dependencies between software modules. Traditional dependency graph construction methods face many challenges when dealing with large-scale software systems, such as high complexity, difficulty in standardization, and difficulty in accurately matching semantics under different programming styles. In the process of automatically generating software, a method that can accurately and efficiently analyze and construct system dependency graphs is needed to assist in software design, development, and maintenance.

[0003] In the prior art, there have been many studies on dependency graph construction, such as system dependency graph construction solutions for program understanding and knowledge-driven program dependency graph construction solutions:

[0004] A. System dependency graph construction solution for program understanding

[0005] This scheme reduces the complexity of program standardization and program matching by processing the three stages of program information extraction, control dependency subgraph construction, and data dependency subgraph construction separately, which is very beneficial for the identification of dependency relationships in automatically generated software.

[0006] However, although the above scheme has a series of advantages in program understanding and analysis, it will ignore certain connections between data dependency and control dependency, affecting the consistency and integrity of system construction, and may result in low analysis efficiency and occupy more computing resources when dealing with large-scale software systems;

[0007] B. Knowledge-driven program dependency graph construction solution

[0008] The main principle of this solution is to build a knowledge graph for the program's dependency packages, system libraries, and interpreters. By collecting a large amount of multi-source heterogeneous data, knowledge is extracted and integrated to build a knowledge graph for a specific domain system. In the program dependency inference stage, the solution infers the multi-level dependencies of the target system based on the domain knowledge graph through program analysis and constraint solving methods.

[0009] Although the above solution greatly improves the accuracy and execution efficiency of program construction, the success rate is still low. This is because knowledge graphs usually need to integrate data from multiple different sources, and the quality of these data varies. Even high-quality data sources may contain errors or biases, which will be passed on to the final knowledge graph. In addition, in some cases, the construction and reasoning process behind the knowledge graph may be opaque, which also reduces the interpretability of the system and makes it difficult to be fully trusted.

[0010] It can be seen that the existing technical solutions are mostly concentrated in specific fields or fail to fully consider the dynamic changes in the needs of automatically generated software. Therefore, there is an urgent need for an intelligent dependency graph construction method based on multidimensional analysis and adaptive optimization, which can not only reduce the complexity of the algorithm and improve the accuracy and efficiency of dependency identification when processing large-scale software systems, but also adapt to the changing needs during the software evolution process, thereby providing strong support for the automatic generation of software. Summary of the invention

[0011] 1. Technical issues to be resolved

[0012] In view of the above-mentioned shortcomings of the prior art, the present invention provides an intelligent dependency graph construction method based on multidimensional analysis and adaptive optimization, which can effectively overcome the defects of the prior art that it is difficult to accurately and efficiently analyze and construct software module dependency graphs.

[0013] (II) Technical solution

[0014] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0015] A method for constructing an intelligent dependency graph based on multidimensional analysis and adaptive optimization comprises the following steps:

[0016] S1, read module data, and construct a preliminary dependency graph by preliminarily parsing the dependency relationships between modules;

[0017] S2, performing a multidimensional analysis on the dependency relationship between each pair of modules in the preliminary dependency graph, and calculating a comprehensive score for each pair of modules based on the multidimensional analysis results;

[0018] S3, adjusting the preliminary dependency graph according to the comprehensive scores of each pair of modules to obtain an adjusted dependency graph;

[0019] S4, performing circular dependency detection on the adjusted dependency graph, and merging the static dependency data and dynamic dependency data of each module;

[0020] S5. Perform adaptive optimization according to the complexity of the adjusted dependency graph, generate a final dependency graph, and save and output the final dependency graph.

[0021] Preferably, S1 reads the module data and constructs a preliminary dependency graph by preliminarily parsing the dependency relationships between the modules, including:

[0022] S11, reading module data from the system log, module description file and database, obtaining the attribute set of each module and the module set M;

[0023] Among them, module M iThe attribute set is A(M i )={name(M i ),type(M i ),func(M i ),perf(M i ),time(M i )},name(M i ) is module M i Name, type(M i ) is module M i The type of func(M i ) is module M i Functional matrix, including module M i The functional dependency strength between other modules, perf(M i ) is module M i The performance matrix of the module M i The performance dependence intensity between other modules, time(M i ) is module M i The time matrix of M i The strength of temporal dependencies with other modules;

[0024] Module set M = {M1, M2, ..., M n1}, n1 is the number of modules in the module set M;

[0025] S12, extracting the dependency relationship between modules from the calling relationship, API calling, configuration files and data between modules to obtain the dependency relationship set R between modules;

[0026] Among them, the dependency set R = {(M i ,M j )}, i,j∈[1,n];

[0027] S13, converting the dependency set R into an adjacency matrix A;

[0028] Among them, if the module M i Depends on module M j , then set the corresponding element in the adjacency matrix A to 1, otherwise set it to 0;

[0029] S14, constructing a preliminary dependency graph G = (M, R) according to the module set M and the dependency set R;

[0030] Among them, the module set M is the node set of the preliminary dependency graph G, and the dependency set R is the edge set of the preliminary dependency graph G.

[0031] Preferably, in S2, a multi-dimensional analysis is performed on the dependency relationship between each pair of modules in the preliminary dependency graph, and a comprehensive score of each pair of modules is calculated according to the multi-dimensional analysis result, including:

[0032] S21. Calculate the module M in the preliminary dependency graph G i With module M j The functional dimension parameter F(M i ,M j ):

[0033]

[0034] Among them, func(M i ) k For module M i The function matrix func(M i ), func(M j ) k For module M j The function matrix func(M j ), n2 is the kth element in the module M i The function matrix func(M i ) and module M j The function matrix func(M j ), ||·||2 means calculating the L2 norm;

[0035] S22. Calculate the module M in the preliminary dependency graph G i With module M j The performance dimension parameter P(M i ,M j ):

[0036]

[0037] Among them, perf(M i ) is module M i The performance matrix, perf(M j ) is module M j Performance matrix;

[0038] S23. Calculate the module M in the preliminary dependency graph G i With module M j The time dimension parameter T(M i ,M j ):

[0039]

[0040] Among them, time(M i ) is module M iThe time matrix, time(M j ) is module M j The time matrix;

[0041] S24. Calculate the module M in the preliminary dependency graph G i With module M j The comprehensive score S(M i ,M j ):

[0042] S(M i ,M j )=ω1·F(M i ,M j )+ω2·P(M i ,M j )+ω3·T(M i ,M j );

[0043] Among them, ω1, ω2, and ω3 are the functional dimension parameters F(M i ,M j ), performance dimension parameter P(M i ,M j ), time dimension parameter T(M i ,M j ), and ω1+ω2+ω3=1.

[0044] Preferably, in S3, the preliminary dependency graph is adjusted according to the comprehensive score of each pair of modules to obtain an adjusted dependency graph, including:

[0045] S31. Determine the module M in the preliminary dependency graph G i With module M j The comprehensive score S(M i ,M j ), the relationship between the preset thresholds, if the comprehensive score S(M i ,M j ) is greater than a preset threshold, the corresponding element in the adjacency matrix A is set to 1 to obtain the updated adjacency matrix A'; otherwise, the adjacency matrix A is directly used as the updated adjacency matrix A';

[0046] S32, updating the dependency set R according to the updated adjacency matrix A' to obtain an updated dependency set R';

[0047] S33, adjusting the preliminary dependency graph G according to the module set M and the updated dependency set R', to obtain an adjusted dependency graph G'=(M, R');

[0048] Among them, the module set M is used as the node set of the adjusted dependency graph G', and the updated dependency set R' is used as the edge set of the adjusted dependency graph G'.

[0049] Preferably, in S4, a circular dependency detection is performed on the adjusted dependency graph, and static dependency data and dynamic dependency data of each module are merged, including:

[0050] S41, using a depth-first search algorithm to perform circular dependency detection on the adjusted dependency graph G', if there is a circular dependency, then optimize the circular dependency and merge the static dependency data and dynamic dependency data of each module, otherwise enter S42;

[0051] S42, checking the dynamic dependencies of each module, and collecting the dynamic dependencies of each module;

[0052] S43, analyzing the impact of the dynamic dependency of each module on the adjustment dependency graph G', and merging the static dependency data and dynamic dependency data of each module.

[0053] Preferably, in S5, adaptive optimization is performed according to the complexity of the adjustment dependency graph to generate a final dependency graph, and the final dependency graph is saved and output, including:

[0054] S51. Calculate the complexity C(G') of adjusting the dependency graph G':

[0055]

[0056] Where n1 is the number of modules in the module set M, n3 is the number of dependencies in the updated dependency set R', deg(M i ) is module M i The node degree of module M i The number of dependencies, A' K For module M i Depends on the Kth element corresponding to other modules in the updated adjacency matrix A', n4 is the Kth element corresponding to module M i The number of modules that depend on other modules;

[0057] S52, judging whether adaptive optimization is needed according to the complexity C(G'), if the complexity C(G') is greater than the preset threshold, adjusting the optimization strategy to minimize the complexity C(G'), and using the adjusted optimization strategy to adaptively optimize the updated dependency set R' to obtain the optimized dependency set R o , and returns to S51 until the complexity C(G') is no greater than the preset threshold, at which time the final dependency set R is obtained final Otherwise, the updated dependency set R' is directly used as the final dependency set R final ;

[0058] S53, according to the module set M and the final dependency set R final , generate the final dependency graph G final =(M,R final );

[0059] Among them, the module set M is used as the final dependency graph G final The node set, the final dependency set R final As the final dependency graph G final The edge set of

[0060] S54, the final dependency graph G final Stored in a structured data format for subsequent visualization and query, and output the final dependency graph G after storage final ;

[0061] Among them, structured data formats include JSON, XML, and graph databases.

[0062] (III) Beneficial effects

[0063] Compared with the prior art, the intelligent dependency graph construction method based on multidimensional analysis and adaptive optimization provided by the present invention can dynamically analyze and process the dependency relationships between modules in large-scale software systems by introducing artificial intelligence, machine learning, multidimensional data analysis, adaptive optimization and visualization technology, thereby improving the system's analysis capabilities and computing efficiency for multidimensional data, and realizing automatic construction and optimization of dependency graphs. Combined with the parallel computing framework and data security technology, the present invention supports dependency analysis of large-scale software systems while ensuring data security, and provides intuitive visualization and interactive functions to help users quickly understand and adjust the dependency structure of the software system. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0065] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] An intelligent dependency graph construction method based on multidimensional analysis and adaptive optimization, such as Figure 1 As shown, S1, reads module data, and constructs a preliminary dependency graph by preliminarily parsing the dependency relationships between modules, specifically including:

[0068] S11, reading module data from the system log, module description file and database, obtaining the attribute set of each module and the module set M;

[0069] Among them, module M i The attribute set is A(M i )={name(M i ),type(M i ),func(M i ),perf(M i ),time(M i )},name(M i ) is module M i Name, type(M i ) is module M i The type of func(M i ) is module M i Functional matrix, including module M i The functional dependency strength between other modules, perf(M i ) is module M i The performance matrix of the module M i The performance dependence intensity between other modules, time(M i ) is module M i The time matrix of M i The strength of temporal dependencies with other modules;

[0070] Module set M = {M1, M2, ..., M n1}, n1 is the number of modules in the module set M;

[0071] S12, extracting the dependency relationship between modules from the calling relationship, API calling, configuration files and data between modules to obtain the dependency relationship set R between modules;

[0072] Among them, the dependency set R = {(M i ,M j )}, i,j∈[1,n];

[0073] S13, converting the dependency set R into an adjacency matrix A;

[0074] Among them, if the module M i Depends on module M j , then set the corresponding element in the adjacency matrix A to 1, otherwise set it to 0;

[0075] S14, constructing a preliminary dependency graph G = (M, R) according to the module set M and the dependency set R;

[0076] Among them, the module set M is the node set of the preliminary dependency graph G, and the dependency set R is the edge set of the preliminary dependency graph G.

[0077] S2. Perform a multi-dimensional analysis on the dependency relationship between each pair of modules in the preliminary dependency graph, and calculate the comprehensive score of each pair of modules based on the multi-dimensional analysis results, including:

[0078] S21. Calculate the module M in the preliminary dependency graph G i With module M j The functional dimension parameter F(M i ,M j ):

[0079]

[0080] Among them, func(M i ) k For module M i The function matrix func(M i ), func(M j ) k For module M j The function matrix func(M j ), n2 is the kth element in the module M i The function matrix func(M i ) and module M j The function matrix func(M j ), ||·||2 means calculating the L2 norm;

[0081] S22. Calculate the module M in the preliminary dependency graph G i With module M j The performance dimension parameter P(M i ,M j ):

[0082]

[0083] Among them, perf(M i ) is module M i The performance matrix, perf(M j ) is module M j Performance matrix;

[0084] S23. Calculate the module M in the preliminary dependency graph G i With module M j The time dimension parameter T(M i ,M j ):

[0085]

[0086] Among them, time(M i ) is module M i The time matrix, time(M j ) is module M j The time matrix;

[0087] S24. Calculate the module M in the preliminary dependency graph G i With module M j The comprehensive score S(M i ,M j ):

[0088] S(M i ,M j )=ω1·F(M i ,M j )+ω2·P(M i ,M j )+ω3·T(M i ,M j );

[0089] Among them, ω1, ω2, and ω3 are the functional dimension parameters F(M i ,M j ), performance dimension parameter P(M i ,M j ), time dimension parameter T(M i ,M j ), and ω1+ω2+ω3=1.

[0090] S3. Adjust the preliminary dependency graph according to the comprehensive score of each pair of modules to obtain an adjusted dependency graph, which specifically includes:

[0091] S31. Determine the module M in the preliminary dependency graph G i With module M j The comprehensive score S(M i ,Mj ), the relationship between the preset thresholds, if the comprehensive score S(M i ,M j ) is greater than a preset threshold, the corresponding element in the adjacency matrix A is set to 1 to obtain the updated adjacency matrix A'; otherwise, the adjacency matrix A is directly used as the updated adjacency matrix A';

[0092] S32, updating the dependency set R according to the updated adjacency matrix A' to obtain an updated dependency set R';

[0093] S33, adjusting the preliminary dependency graph G according to the module set M and the updated dependency set R', to obtain an adjusted dependency graph G'=(M, R');

[0094] Among them, the module set M is used as the node set of the adjusted dependency graph G', and the updated dependency set R' is used as the edge set of the adjusted dependency graph G'.

[0095] S4. Perform circular dependency detection on the adjusted dependency graph and merge the static dependency data and dynamic dependency data of each module, specifically including:

[0096] S41, using a depth-first search algorithm to perform circular dependency detection on the adjusted dependency graph G', if there is a circular dependency, then optimize the circular dependency and merge the static dependency data and dynamic dependency data of each module, otherwise enter S42;

[0097] S42, checking the dynamic dependencies of each module, and collecting the dynamic dependencies of each module;

[0098] S43, analyzing the impact of the dynamic dependency of each module on the adjustment dependency graph G', and merging the static dependency data and dynamic dependency data of each module.

[0099] S5. Adaptively optimize according to the complexity of the dependency graph, generate the final dependency graph, and save and output the final dependency graph, including:

[0100] S51. Calculate the complexity C(G') of adjusting the dependency graph G':

[0101]

[0102] Where n1 is the number of modules in the module set M, n3 is the number of dependencies in the updated dependency set R', deg(M i ) is module M i The node degree of module M i The number of dependencies, A' K For module M iDepends on the Kth element corresponding to other modules in the updated adjacency matrix A', n4 is the Kth element corresponding to module M i The number of modules that depend on other modules;

[0103] S52, judging whether adaptive optimization is needed according to the complexity C(G'), if the complexity C(G') is greater than the preset threshold, adjusting the optimization strategy to minimize the complexity C(G'), and using the adjusted optimization strategy to adaptively optimize the updated dependency set R' to obtain the optimized dependency set R o , and returns to S51 until the complexity C(G') is no greater than the preset threshold, at which time the final dependency set R is obtained final Otherwise, the updated dependency set R' is directly used as the final dependency set R final ;

[0104] S53, according to the module set M and the final dependency set R final , generate the final dependency graph G final =(M,R final );

[0105] Among them, the module set M is used as the final dependency graph G final The node set, the final dependency set R final As the final dependency graph G final The edge set of

[0106] S54, the final dependency graph G final Stored in a structured data format for subsequent visualization and query, and output the final dependency graph G after storage final ;

[0107] Among them, structured data formats include JSON, XML, and graph databases.

[0108] In the technical solution of this application, by introducing artificial intelligence, machine learning, multidimensional data analysis, adaptive optimization and visualization technology, it is possible to dynamically analyze and process the dependencies between modules in large-scale software systems, improve the system's analysis capabilities and computing efficiency for multidimensional data, and realize the automatic construction and optimization of dependency graphs. Combined with the parallel computing framework and data security technology, the present invention supports dependency analysis of large-scale software systems while ensuring data security, and provides intuitive visualization and interactive functions to help users quickly understand and adjust the dependency structure of the software system.

[0109] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for constructing an intelligent dependency graph based on multidimensional analysis and adaptive optimization, characterized in that: The following steps are involved: S1, read module data, and construct a preliminary dependency graph by preliminarily parsing the dependency relationships between modules; S2, performing a multidimensional analysis on the dependency relationship between each pair of modules in the preliminary dependency graph, and calculating a comprehensive score for each pair of modules based on the multidimensional analysis results; S3, adjusting the preliminary dependency graph according to the comprehensive scores of each pair of modules to obtain an adjusted dependency graph; S4, performing circular dependency detection on the adjusted dependency graph, and merging the static dependency data and dynamic dependency data of each module; S5. Perform adaptive optimization according to the complexity of the adjusted dependency graph, generate a final dependency graph, and save and output the final dependency graph.

2. The intelligent dependency graph construction method based on multidimensional analysis and adaptive optimization according to claim 1 is characterized in that: S1 reads the module data and constructs a preliminary dependency graph by preliminarily parsing the dependencies between modules, including: S11, reading module data from the system log, module description file and database, obtaining the attribute set of each module and the module set M; Among them, module M i The attribute set is A(M i )={name(M i ),type(M i ),func(M i ),perf(M i ),time(M i )},name(M i ) is module M i Name, type(M i ) is module M i The type of func(M i ) is module M i The functional matrix contains modules M i The functional dependency strength between other modules, perf(M i ) is module M i The performance matrix of the module M i The performance dependence intensity between other modules, time(M i ) is module M i The time matrix of M i The strength of temporal dependencies with other modules; Module set M = {M1, M2, ..., M n1 }, n1 is the number of modules in the module set M; S12, extracting the dependency relationship between modules from the calling relationship, API calling, configuration files and data between modules to obtain the dependency relationship set R between modules; Among them, the dependency set R = {(M i ,M j )}, i,j∈[1,n]; S13, converting the dependency set R into an adjacency matrix A; Among them, if the module M i Depends on module M j , then set the corresponding element in the adjacency matrix A to 1, otherwise set it to 0; S14, constructing a preliminary dependency graph G = (M, R) according to the module set M and the dependency set R; Among them, the module set M is the node set of the preliminary dependency graph G, and the dependency set R is the edge set of the preliminary dependency graph G.

3. The intelligent dependency graph construction method based on multidimensional analysis and adaptive optimization according to claim 2 is characterized in that: S2 performs a multidimensional analysis of the dependency relationship between each pair of modules in the preliminary dependency graph, and calculates the comprehensive score of each pair of modules based on the multidimensional analysis results, including: S21. Calculate the module M in the preliminary dependency graph G i With module M j The functional dimension parameter F(M i ,M j ): Among them, func(M i ) k For module M i The function matrix func(M i ), func(M j ) k For module M j The function matrix func(M j ), n2 is the kth element in the module M i The function matrix func(M i ) and module M j The function matrix func(M j ), ||·||2 means calculating the L2 norm; S22. Calculate the module M in the preliminary dependency graph G i With module M j The performance dimension parameter P(M i ,M j ): Among them, perf(M i ) is module M i The performance matrix, perf(M j ) is module M j Performance matrix; S23. Calculate the module M in the preliminary dependency graph G i With module M j The time dimension parameter T(M i ,M j ): Among them, time(M i ) is module M i The time matrix, time(M j ) is module M j The time matrix; S24. Calculate the module M in the preliminary dependency graph G i With module M j The comprehensive score S(M i ,M j ): S(M i ,M j )=ω1·F(M i ,M j )+ω2·P(M i ,M j )+ω3·T(M i ,M j ); Among them, ω1, ω2, and ω3 are the functional dimension parameters F(M i ,M j ), performance dimension parameter P(M i ,M j ), time dimension parameter T(M i ,M j ), and ω1+ω2+ω3=1.

4. The intelligent dependency graph construction method based on multidimensional analysis and adaptive optimization according to claim 3 is characterized in that: In S3, the preliminary dependency graph is adjusted according to the comprehensive score of each pair of modules to obtain an adjusted dependency graph, including: S31. Determine the module M in the preliminary dependency graph G i With module M j The comprehensive score S(M i ,M j ), the relationship between the preset thresholds, if the comprehensive score S(M i ,M j ) is greater than a preset threshold, the corresponding element in the adjacency matrix A is set to 1 to obtain the updated adjacency matrix A'; otherwise, the adjacency matrix A is directly used as the updated adjacency matrix A'; S32, updating the dependency set R according to the updated adjacency matrix A' to obtain an updated dependency set R'; S33, adjusting the preliminary dependency graph G according to the module set M and the updated dependency set R', to obtain an adjusted dependency graph G'=(M, R'); Among them, the module set M is used as the node set of the adjusted dependency graph G', and the updated dependency set R' is used as the edge set of the adjusted dependency graph G'.

5. The intelligent dependency graph construction method based on multidimensional analysis and adaptive optimization according to claim 4 is characterized in that: S4 performs circular dependency detection on the adjusted dependency graph and merges the static dependency data and dynamic dependency data of each module, including: S41, using a depth-first search algorithm to perform circular dependency detection on the adjusted dependency graph G', if there is a circular dependency, then optimize the circular dependency and merge the static dependency data and dynamic dependency data of each module, otherwise enter S42; S42, checking the dynamic dependencies of each module, and collecting the dynamic dependencies of each module; S43, analyzing the impact of the dynamic dependency of each module on the adjustment dependency graph G', and merging the static dependency data and dynamic dependency data of each module.

6. The intelligent dependency graph construction method based on multidimensional analysis and adaptive optimization according to claim 5 is characterized in that: In S5, adaptive optimization is performed according to the complexity of the dependency graph, the final dependency graph is generated, and the final dependency graph is saved and output, including: S51. Calculate the complexity C(G') of adjusting the dependency graph G': Where n1 is the number of modules in the module set M, n3 is the number of dependencies in the updated dependency set R', deg(M i ) is module M i The node degree of module M i The number of dependencies, A' K For module M i Depends on the Kth element corresponding to other modules in the updated adjacency matrix A', n4 is the Kth element corresponding to module M i The number of modules that depend on other modules; S52, judging whether adaptive optimization is needed according to the complexity C(G'), if the complexity C(G') is greater than the preset threshold, adjusting the optimization strategy to minimize the complexity C(G'), and using the adjusted optimization strategy to adaptively optimize the updated dependency set R' to obtain the optimized dependency set R o , and returns to S51 until the complexity C(G') is no greater than the preset threshold, at which time the final dependency set R is obtained final Otherwise, the updated dependency set R' is directly used as the final dependency set R final ; S53, according to the module set M and the final dependency set R final , generate the final dependency graph G final =(M,R final ); Among them, the module set M is used as the final dependency graph G final The node set, the final dependency set R final As the final dependency graph G final The edge set of S54, the final dependency graph G final Stored in a structured data format for subsequent visualization and query, and output the final dependency graph G after storage final ; Among them, structured data formats include JSON, XML, and graph databases.

Images