A method and system for quantifying defect severity based on propagation characteristics
By constructing a dynamic dependency graph and combining it with a multi-dimensional quantification method of propagation topology complexity, depth and strength, the problem of insufficient accuracy in defect assessment under complex systems is solved, realizing automated and dynamic defect severity assessment, and improving the accuracy and efficiency of assessment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU BAOLUN ELECTRONICS CO LTD
- Filing Date
- 2026-04-07
- Publication Date
- 2026-07-21
AI Technical Summary
Existing technologies lack multi-dimensional modeling of anomaly propagation when assessing the severity of software defects, making it difficult to quantify the scope of defect impact in complex systems. This results in insufficient accuracy and timeliness of the assessment, failing to meet the actual needs of large-scale, highly complex software systems.
By collecting execution trajectory data, a dynamic dependency graph is constructed to represent the probability of anomaly propagation between modules. The propagation topology complexity, depth and intensity are combined for multi-dimensional quantification, and the anomaly propagation path is dynamically tracked. A dynamic weight adjustment mechanism is used to optimize the evaluation indicators.
It enables precise quantification of abnormal propagation characteristics, improves the accuracy and objectivity of defect severity assessment, reduces subjective judgment errors, and enhances assessment efficiency and response speed.
Smart Images
Figure CN122432031A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of computer software technology, and in particular relates to a method and system for measuring the severity of defects based on propagation characteristics. Background Technology
[0002] In recent years, with the widespread deployment of distributed software, microservice architecture, and large-scale business systems, the number of software functional modules has continued to increase, and service call chains have become increasingly longer. This has led to increasingly stringent requirements for defect identification, fault location, and stability assurance. However, current software defect severity assessment schemes still rely on manual experience and simple rule-based judgments. The scope of defect impact in complex systems is difficult to quantify automatically, and the objectivity and accuracy of the assessments are insufficient, failing to meet the actual quality control needs of large-scale, highly complex software systems.
[0003] Most current defect assessment methods focus on calculating single-dimensional indicators, paying attention only to the length of the abnormal path or the number of affected nodes, while ignoring the propagation depth, probability, and topological complexity of the anomaly within the system. This limits their application to scenarios with simple module relationships and straightforward business logic. When applied to real-world distributed systems, they are constrained by factors such as ambiguous propagation paths, insufficient quantification dimensions, and assessment lag, failing to achieve accurate and dynamic defect severity determination. In other words, the core of defect severity quantification in complex systems lies in achieving multi-dimensional modeling and real-time calculation of anomaly propagation characteristics. However, existing solutions suffer from incomplete propagation process characterization, unreasonable influence weights, and weak dynamic adaptability, making it difficult to meet the accuracy and timeliness requirements of practical engineering. Summary of the Invention
[0004] This application proposes a defect severity quantification method and system based on propagation characteristics. By collecting execution trajectories to construct a dynamic dependency graph, it accurately characterizes the probability of abnormal propagation between modules and can automatically track abnormal propagation paths. Combining propagation topology complexity, propagation depth, and intensity to achieve multi-dimensional quantification, it breaks through the limitations of traditional single-dimensional assessment and significantly improves the accuracy and objectivity of defect severity determination.
[0005] This invention provides a method for measuring defect severity based on propagation characteristics, comprising:
[0006] Collect execution trajectory data of the software system during runtime;
[0007] A dynamic dependency graph is constructed based on the execution trajectory data, and the directed edge weights of the dynamic dependency graph are used to characterize the probability of anomaly propagation.
[0008] In response to an abnormal event, the starting node for an abnormal propagation is determined in the dynamic dependency graph, and the dynamic dependency graph is traversed to generate one or more abnormal propagation paths;
[0009] The propagation topology complexity, propagation depth, and propagation intensity of the abnormal propagation path are calculated to obtain the defect severity of the abnormal event.
[0010] According to the present invention, a defect severity quantification method based on propagation characteristics is provided, wherein the weight of the directed edge of the dynamic dependency graph is determined according to the calling frequency between modules or functions, specifically: the ratio of the number of times a module calls another module to the total number of times the module is called is used as the probability of an anomaly propagating along the edge.
[0011] According to the present invention, a defect severity quantification method based on propagation characteristics is provided, wherein the propagation topology complexity is determined by combining the edge density of the anomalous propagation subgraph with the degree of dispersion of the propagation path length.
[0012] According to the present invention, a method for measuring the severity of defects based on propagation characteristics is provided, wherein the propagation depth is obtained by weighted summation of propagation path length, number of affected nodes, and propagation entropy;
[0013] The propagation intensity is obtained by multiplying the propagation probabilities between adjacent nodes on the abnormal propagation path, and the system propagation intensity is the maximum value of the propagation intensity of all propagation paths.
[0014] The defect severity quantification method based on propagation characteristics provided by the present invention further includes:
[0015] Calculate the system impact coefficient, which is determined based on the preset business weight of the module affected by the anomaly and the structural centrality of the module in the dynamic dependency graph;
[0016] Calculate the time decay factor, which is used to characterize the decay characteristics of the abnormal influence over time.
[0017] According to the present invention, a defect severity quantification method based on propagation characteristics is provided, wherein the system impact coefficient is obtained by multiplying the business weights of each module affected by the anomaly with their corresponding structural centrality and then summing them up;
[0018] The time decay factor is determined using an exponential decay model.
[0019] According to the present invention, a defect severity quantification method based on propagation characteristics is provided, wherein the dynamic weight adjustment mechanism optimizes the weight of each indicator by minimizing the sum of squared errors between the predicted severity and the actual severity.
[0020] The severity of the defect is obtained by multiplying the weighted sum of propagation depth, system influence coefficient, and propagation topology complexity by a time decay factor.
[0021] The present invention also provides a defect severity quantification system based on propagation characteristics, the system comprising:
[0022] The data acquisition module is used to collect execution trajectory data during the operation of the software system;
[0023] The graph construction module is used to construct a dynamic dependency graph based on the execution trajectory data, wherein the directed edge weights of the dynamic dependency graph are used to characterize the probability of anomaly propagation.
[0024] The path analysis module is used to determine the starting node of an anomaly propagation in the dynamic dependency graph in response to an anomaly event, and to traverse the dynamic dependency graph to generate one or more anomaly propagation paths.
[0025] The feature calculation module is used to calculate the propagation topology complexity, propagation depth, and propagation strength of the abnormal propagation path;
[0026] An evaluation module is used to determine the severity of the defect in the anomalous event based on the propagation topology complexity, propagation depth, and propagation intensity.
[0027] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the defect severity quantification method based on propagation characteristics as described above.
[0028] This invention provides a method and system for quantifying defect severity based on propagation characteristics. By collecting execution trajectory data of the software system in real time and constructing a dynamic dependency graph, it can accurately characterize the calling relationships between modules and the probability of anomaly propagation, providing a real and dynamic data foundation for anomaly propagation analysis. Compared with traditional methods that rely on static structures and log analysis, this method is more closely aligned with the actual operating state of the system. By locating the anomaly initiation node and traversing to generate a complete propagation path, the spread process of the anomaly within the system can be fully reconstructed, effectively improving the completeness and accuracy of path identification. By introducing three core indicators—propagation topology complexity, propagation depth, and propagation intensity—multi-dimensional quantitative calculation of anomaly propagation characteristics is achieved, solving the one-sidedness problem caused by existing technologies that rely solely on single-dimensional assessments such as path length and the number of affected nodes, significantly improving the accuracy and objectivity of defect severity determination. Furthermore, this solution can be fully automated, requiring no human intervention, reducing subjective judgment errors, improving defect assessment efficiency and response speed, and quickly and dynamically reflecting the actual impact and severity of defects on the system. Attached Figure Description
[0029] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0030] Figure 1 This is a flowchart illustrating a defect severity quantification method based on propagation characteristics provided in an embodiment of this application;
[0031] Figure 2 This is a schematic diagram of a defect severity quantification system based on propagation characteristics provided in an embodiment of this application. Detailed Implementation
[0033] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.
[0034] Example 1
[0035] To address the problems in existing technologies, this invention proposes a defect severity quantification method and system based on propagation characteristics. By collecting execution trajectories to construct a dynamic dependency graph, it accurately characterizes the probability of anomaly propagation between modules and can automatically track anomaly propagation paths. Combining propagation topology complexity, propagation depth, and intensity to achieve multi-dimensional quantification overcomes the limitations of traditional single-dimensional assessment, significantly improving the accuracy and objectivity of defect severity determination. Figure 1 As shown, including but not limited to the following steps:
[0036] S1: Collect execution trajectory data of the software system during runtime.
[0037] Step S1 forms the data foundation for the entire dynamic defect severity quantification method. It uses an event-driven dynamic tracing mechanism to monitor the software system's operation in real time, collecting data including function or service call relationships, call time sequences, exception event records, and variable state changes. When an exception occurs, it automatically records the complete call stack information and converts it into exception propagation path data. The collected execution trajectory data is preprocessed, including filtering irrelevant logs, unifying module identifiers, and constructing call sequences, providing standardized and reliable foundational data for subsequent dependency graph construction and propagation analysis.
[0038] Furthermore, the dynamic tracing layer employs an event-driven data collection mechanism to monitor key behaviors during system operation in real time. When an anomaly occurs, it automatically records the complete call stack information and transforms it into anomaly propagation path data. Therefore, the runtime dynamic data collection method can accurately reflect the actual call behavior of the system, which is more suitable for the complex call scenarios of microservice architectures compared to static analysis, ensuring the authenticity and timeliness of subsequent analysis.
[0039] S2: Construct a dynamic dependency graph based on the execution trajectory data, wherein the directed edge weights of the dynamic dependency graph are used to characterize the probability of anomaly propagation.
[0040] As a further optional embodiment, the weight of the directed edge in the dynamic dependency graph is determined based on the calling frequency between modules or functions. Specifically, the ratio of the number of times a module calls another module to the total number of times the module is called is used as the probability of an anomaly propagating along the edge.
[0041] In step S2, a system dynamic dependency graph model is constructed based on the preprocessed execution trajectory data. In the model, system modules or functions are used as nodes, inter-module call relationships are used as directed edges, and the inter-module call frequency is used to calculate the edge weight. This weight directly represents the probability of an anomaly propagating along the corresponding call relationship.
[0042] Furthermore, dynamic dependency graphs can be updated in real time as the system runs, accurately reflecting the actual dependency strength between modules. Compared with static structure graphs, they can better reflect the true propagation trend of anomalies and improve the accuracy of path analysis and quantitative calculation.
[0043] The edge weights are calculated based on the frequency of calls between modules as follows:
[0044] ;
[0045] in: This indicates the number of times module A calls module B. This represents the total number of times module A is called.
[0046] S3: In response to an abnormal event, determine the starting node of the abnormal propagation in the dynamic dependency graph, and traverse the dynamic dependency graph to generate one or more abnormal propagation paths.
[0047] In step S3, when an abnormal event is detected, the starting node of the abnormal propagation is located based on the abnormal call stack information; the dynamic dependency graph is traversed using a breadth-first search or depth-first search algorithm, all nodes that the abnormal can propagate along the directed edges are traced, the set of nodes that the abnormal may affect is identified, and one or more complete abnormal propagation paths are generated.
[0048] Furthermore, based on the full path traversal of the dynamic dependency graph, the propagation range of the exception in the complex call chain can be fully restored, avoiding the problems of incomplete path and underestimated impact range caused by relying solely on call stack interception, and ensuring that the identification of exception impact is comprehensive and without omission.
[0049] Specifically, when an abnormal event is detected, the starting point of the abnormal propagation is determined based on the call stack information, and the abnormal propagation path is traced in the dynamic dependency graph.
[0050] The abnormal propagation path is represented as:
[0051] ;
[0052] in: This is the starting node of the anomaly. This is the termination node for propagation.
[0053] By using graph traversal algorithms (such as breadth-first search or depth-first search), the set of nodes that the anomaly may affect can be identified, thus obtaining the complete set of propagation paths.
[0054] ;
[0055] This set of paths is used for subsequent propagation feature calculations.
[0056] To achieve complete identification of the abnormal propagation range, this invention provides the following example code:
[0057] functionanalyzePropagation(graph,startNode):
[0058] # Initialize the queue for hierarchical traversal of the dependency graph.
[0059] #startNode represents the starting node where the exception occurred.
[0060] queue=[startNode]
[0061] # Used to record visited nodes to avoid repeated traversal or loops.
[0062] visited = set()
[0063] #Continue iterating through the queue as long as it is not empty.
[0064] while queue not empty:
[0065] # Retrieve the current node from the queue
[0066] #pop() means retrieving a node from the queue for processing.
[0067] node = queue.pop()
[0068] # Mark the current node as visited
[0069] visited.add(node)
[0070] # Traverse all adjacent nodes of the current node
[0071] #neighbors(node) represents nodes in the dependency graph that have a calling relationship with this node.
[0072] forneighboringraph.neighbors(node):
[0073] #If the adjacent node has not yet been visited
[0074] ifneighbornotinvisited:
[0075] #Add this node to the queue
[0076] # indicates that the node may be affected by abnormal propagation.
[0077] queue.push(neighbor)
[0078] # Returns the set of all visited nodes
[0079] #That is, all nodes that can be propagated from startNode.
[0080] returnvisited
[0081] S4: Calculate the propagation topology complexity, propagation depth, and propagation intensity of the abnormal propagation path to obtain the defect severity of the abnormal event.
[0082] As a further optional embodiment, the propagation topology complexity is determined by combining the edge density of the anomalous propagation subgraph with the degree of dispersion of the propagation path length.
[0083] As a further optional embodiment, the propagation depth is obtained by weighted summation of the propagation path length, the number of affected nodes, and the propagation entropy;
[0084] The propagation intensity is obtained by multiplying the propagation probabilities between adjacent nodes on the abnormal propagation path, and the system propagation intensity is the maximum value of the propagation intensity of all propagation paths.
[0085] As a further optional embodiment, it also includes: calculating a system impact coefficient, the system impact coefficient being determined based on the preset business weight of the module affected by the anomaly and the structural centrality of the module in the dynamic dependency graph;
[0086] Calculate the time decay factor, which is used to characterize the decay characteristics of the abnormal influence over time.
[0087] Building upon the original approach, a system impact coefficient and a time decay factor are further introduced. The system impact coefficient is determined based on the preset business weights of the affected modules and the centrality of the node structure. The time decay factor uses an exponential decay model to characterize the decay characteristics of the defect's impact over time. Simultaneously, through a dynamic weight adjustment mechanism, the weights of each indicator are optimized with the goal of minimizing the sum of squared errors between the predicted severity and the actual severity. The propagation depth, system impact coefficient, and propagation topology complexity are weighted and summed, and then multiplied by the time decay factor to obtain the final defect severity, which is then divided into four levels: fatal, severe, moderate, and minor, based on a threshold.
[0088] As a further optional embodiment, the system impact coefficient is obtained by multiplying the business weights of each module affected by the anomaly by their corresponding structural centrality and then summing them up.
[0089] The time decay factor is determined using an exponential decay model.
[0090] As a further optional embodiment, the dynamic weight adjustment mechanism optimizes the weights of each indicator by minimizing the sum of squared errors between the predicted severity and the actual severity.
[0091] The severity of the defect is obtained by multiplying the weighted sum of propagation depth, system influence coefficient, and propagation topology complexity by a time decay factor.
[0092] 4.1 Calculation of Propagation Depth
[0093] Calculate the propagation depth of the anomaly propagation path. Propagation depth is a core indicator for measuring the scope and complexity of anomaly propagation. A multi-dimensional propagation depth quantification method is used, including path length, number of affected nodes, and propagation entropy.
[0094] 1. Path length metric
[0095] This indicates the maximum number of nodes in the anomaly propagation path: ;
[0096] 2. Number of affected nodes
[0097] This indicates the total number of nodes affected by the anomaly: ;
[0098] 3. Propagation Entropy
[0099] Propagation entropy is used to measure the uncertainty of a propagation path.
[0100] First, calculate the propagation probability for each propagation path: ;
[0101] in This represents the path propagation intensity.
[0102] Propagation entropy is defined as: ;
[0103] in This represents the probability of node propagation.
[0104] The formula for calculating the overall propagation depth is: ;
[0105] in , , Weights for the propagation dimension.
[0106] 4.2 Calculation of Abnormal Propagation Intensity
[0107] To further describe the level of risk of abnormal transmission, a transmission intensity index is introduced.
[0108] Let the propagation path be: ;
[0109] The path propagation strength is defined as: ;
[0110] in: This represents the propagation probability between nodes.
[0111] The system propagation strength is defined as: ;
[0112] The greater the propagation intensity, the easier it is for the anomaly to spread along that path.
[0113] 4.3 Calculation of Propagation Topological Complexity
[0114] To reflect the complexity of the anomaly propagation structure, a propagation topology complexity index is introduced.
[0115] The propagation topology complexity is defined as: ;
[0116] in: To propagate the number of edges in the subgraph, To propagate the number of nodes in the subgraph, This represents the standard deviation of the propagation path length.
[0117] This metric measures the complexity of the propagation structure. Higher complexity indicates a more complex anomaly propagation structure.
[0118] 4.4 Calculation of System Influence Coefficient
[0119] To reflect the degree of impact of anomalies on the core modules of the system, a system impact coefficient S is introduced.
[0120] This indicator takes into account: module level weight and node centrality.
[0121] Module hierarchy weights reflect the module's business importance within the system. For example:
[0122] Core business modules 0.7 Important service modules 0.5 Auxiliary module 0.3
[0123] Table 1
[0124] Node centrality can be calculated using graph algorithms, such as PageRank or betweenness centrality.
[0125] The formula for calculating the system influence coefficient is: ;
[0126] in: For module-level weights, Let be the node centrality.
[0127] 4.5 Calculation of Time Decay Factor
[0128] Considering that the impact of defects gradually weakens over time, a time decay model is introduced, which can dynamically reflect the time-dependent impact of defects.
[0129] The time decay factor is defined as: ;
[0130] in: The attenuation rate, This is the time difference between the time the defect occurred and the current time.
[0131] 4.6 Dynamic Weight Adjustment Mechanism
[0132] To improve the accuracy of severity assessment, a dynamic weight adjustment mechanism is designed.
[0133] Based on historical defect data, the weights of each indicator are automatically adjusted through an optimization algorithm to minimize the error between the predicted severity and the actual severity.
[0134] The weights satisfy: ;
[0135] The optimization objective is: ;
[0136] 4.7 Defect Severity Calculation and Grading
[0137] The final defect severity is calculated using the following formula:
[0138] ;
[0139] Combined with time decay: ;
[0140] Defect levels are classified based on the calculation results:
[0141] ≥8 fatal 5-8 serious 3-5 generally <3 slight
[0142] Table 2
[0143] This invention provides a defect severity quantification method based on propagation characteristics. By collecting execution trajectory data of the software system in real time and constructing a dynamic dependency graph, it can accurately characterize the call relationships between modules and the probability of anomaly propagation, providing a real and dynamic data foundation for anomaly propagation analysis. Compared with traditional methods that rely on static structures and log analysis, this method is more in line with the actual operating state of the system. By locating the anomaly initiation node and traversing to generate a complete propagation path, the spread process of the anomaly within the system can be fully reconstructed, effectively improving the completeness and accuracy of path identification. By introducing three core indicators—propagation topology complexity, propagation depth, and propagation intensity—multi-dimensional quantitative calculation of anomaly propagation characteristics is achieved, solving the one-sidedness problem caused by existing technologies that rely solely on single-dimensional assessments such as path length and the number of affected nodes, and significantly improving the accuracy and objectivity of defect severity determination. Furthermore, this solution can be fully automated, requiring no human intervention, reducing subjective judgment errors, improving defect assessment efficiency and response speed, and quickly and dynamically reflecting the actual impact and severity of defects on the system.
[0144] Example 2
[0145] The following describes a defect severity quantification system based on propagation characteristics provided by the present invention, such as... Figure 2 As shown, the defect severity measurement system based on propagation characteristics described below can be referred to in correspondence with the defect severity measurement method based on propagation characteristics described above.
[0146] A defect severity quantification system based on propagation characteristics, the system comprising:
[0147] The data acquisition module is used to collect execution trajectory data during the operation of the software system;
[0148] The graph construction module is used to construct a dynamic dependency graph based on the execution trajectory data, wherein the directed edge weights of the dynamic dependency graph are used to characterize the probability of anomaly propagation.
[0149] The path analysis module is used to determine the starting node of an anomaly propagation in the dynamic dependency graph in response to an anomaly event, and to traverse the dynamic dependency graph to generate one or more anomaly propagation paths.
[0150] The feature calculation module is used to calculate the propagation topology complexity, propagation depth, and propagation strength of the abnormal propagation path;
[0151] An evaluation module is used to determine the severity of the defect in the anomalous event based on the propagation topology complexity, propagation depth, and propagation intensity.
[0152] This invention provides a defect severity quantification system based on propagation characteristics. By collecting execution trajectory data of the software system in real time and constructing a dynamic dependency graph, it can accurately characterize the call relationships between modules and the probability of anomaly propagation, providing a real and dynamic data foundation for anomaly propagation analysis. Compared with traditional methods that rely on static structures and log analysis, this system is more closely aligned with the actual operating state of the system. By locating the anomaly initiation node and traversing to generate a complete propagation path, the system can comprehensively reconstruct the anomaly's diffusion process within the system, effectively improving the completeness and accuracy of path identification. By introducing three core indicators—propagation topology complexity, propagation depth, and propagation intensity—the system achieves multi-dimensional quantitative calculation of anomaly propagation characteristics, solving the one-sidedness problem caused by existing technologies that rely solely on single-dimensional assessments such as path length and the number of affected nodes. This significantly improves the accuracy and objectivity of defect severity determination. Furthermore, this solution can be fully automated, requiring no human intervention, reducing subjective judgment errors, improving defect assessment efficiency and response speed, and quickly and dynamically reflecting the actual impact and severity of defects on the system.
[0153] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the aforementioned method for measuring defect severity based on propagation characteristics.
[0154] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0155] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do 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 measuring defect severity based on propagation characteristics, characterized in that, include: Collect execution trajectory data of the software system during runtime; A dynamic dependency graph is constructed based on the execution trajectory data, and the directed edge weights of the dynamic dependency graph are used to characterize the probability of anomaly propagation. In response to an abnormal event, the starting node for an abnormal propagation is determined in the dynamic dependency graph, and the dynamic dependency graph is traversed to generate one or more abnormal propagation paths; The propagation topology complexity, propagation depth, and propagation intensity of the abnormal propagation path are calculated to obtain the defect severity of the abnormal event.
2. The method according to claim 1, characterized in that, The weights of the directed edges in the dynamic dependency graph are determined based on the calling frequency between modules or functions. Specifically, the ratio of the number of times a module calls another module to the total number of times the module is called is used as the probability that an anomaly will propagate along that edge.
3. The method according to claim 1, characterized in that, The propagation topology complexity is determined by combining the edge density of the anomalous propagation subgraph with the degree of dispersion of the propagation path length.
4. The method according to claim 1, characterized in that, The propagation depth is obtained by weighted summation of the propagation path length, the number of affected nodes, and the propagation entropy; The propagation intensity is obtained by multiplying the propagation probabilities between adjacent nodes on the abnormal propagation path, and the system propagation intensity is the maximum value of the propagation intensity of all propagation paths.
5. The method according to claim 1, characterized in that, Also includes: Calculate the system impact coefficient, which is determined based on the preset business weight of the module affected by the anomaly and the structural centrality of the module in the dynamic dependency graph; Calculate the time decay factor, which is used to characterize the decay characteristics of the abnormal influence over time.
6. The method according to claim 5, characterized in that, The system impact coefficient is obtained by multiplying the business weights of each module affected by the anomaly by their corresponding structural centrality and then summing the results. The time decay factor is determined using an exponential decay model.
7. The method according to claim 5, characterized in that, The severity of the defect is calculated by weighting the propagation depth, system influence coefficient, propagation topology complexity, and time decay factor. The weights of each indicator are determined through a dynamic weight adjustment mechanism, which uses historical defect data to optimize the weights and minimize the error between the assessment results and the historical true severity.
8. The method according to claim 7, characterized in that, The dynamic weight adjustment mechanism optimizes the weights of each indicator by minimizing the sum of squared errors between the predicted severity and the actual severity. The severity of the defect is obtained by multiplying the weighted sum of propagation depth, system influence coefficient, and propagation topology complexity by a time decay factor.
9. A dynamic quantification system for defect severity based on multidimensional propagation characteristics of execution trajectory, characterized in that, include: The data acquisition module is used to collect execution trajectory data during the operation of the software system; The graph construction module is used to construct a dynamic dependency graph based on the execution trajectory data, wherein the directed edge weights of the dynamic dependency graph are used to characterize the probability of anomaly propagation. The path analysis module is used to determine the starting node of an anomaly propagation in the dynamic dependency graph in response to an anomaly event, and to traverse the dynamic dependency graph to generate one or more anomaly propagation paths. The feature calculation module is used to calculate the propagation topology complexity, propagation depth, and propagation strength of the abnormal propagation path; An evaluation module is used to determine the severity of the defect in the anomalous event based on the propagation topology complexity, propagation depth, and propagation intensity.
10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the defect severity quantification method based on propagation characteristics as described in any one of claims 1 to 8.