Artificial intelligence-based software configuration development management method and device, and storage medium

Abstract

The present application relates to the technical field of data processing, and more particularly to a software configuration development management method and device based on artificial intelligence and a storage medium. The method comprises: extracting a configuration change sequence in a database and performing feature processing; mining a sequence pattern of the configuration change sequence to construct a configuration evolution pattern library; performing intelligent recommendation and conflict prediction of changes based on the sequence pattern; determining system configuration entropy according to the configuration evolution pattern library, and dividing the change risk level. The present application overcomes the defects of the existing configuration management technology, such as insufficient perception of the dynamic evolution relationship between configurations, lack of quantitative control of the overall confusion degree of the configuration change set, and provides a configuration management that can automatically learn the configuration evolution pattern and perform intelligent recommendation of changes and system entropy control based on the pattern.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a software configuration development and management method, apparatus and storage medium based on artificial intelligence. Background Technology

[0002] As the scale and complexity of software systems continue to increase, the management and maintenance of software configurations have become critical aspects affecting system stability and development efficiency. Traditional configuration management relies heavily on manual experience and static rules, making it difficult to adapt to high-frequency, large-scale change requirements, and lacking the ability to automatically and intelligently perceive potential correlations and risks between changes. Against this backdrop, how to leverage artificial intelligence technology to model, predict, and manage the changes in software configurations has become an important research direction for improving system maintainability and security.

[0003] Existing software configuration management methods largely rely on human experience, static rules, or simple version comparisons, lacking the ability to deeply analyze the implicit temporal relationships and causal connections within configuration change sequences. This leads to change recommendations often being based on simple statistics or fixed templates, failing to accurately predict the interconnected changes of related configuration items and easily overlooking potential systemic risks. Furthermore, existing solutions generally lack systematic quantitative assessment methods for the overall stability of the configuration structure and the risks of change operations. Risk control largely depends on manual judgment or post-event failure review, making it difficult to achieve real-time risk identification and hierarchical control before and during changes, thus affecting the continuous reliable evolution of system configurations and operational efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide a software configuration development management method, apparatus, and storage medium based on artificial intelligence, so as to solve at least one of the problems existing in the prior art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] An artificial intelligence-based software configuration development management method includes:

[0007] Extract the configuration change sequence from the database and perform feature processing;

[0008] Discover sequence patterns in configuration change sequences to build a configuration evolution pattern library;

[0009] Intelligent change recommendation and conflict prediction based on sequence patterns;

[0010] The system configuration entropy is determined based on the configuration evolution pattern library, and the change risk level is classified.

[0011] Furthermore, the aggregation time window parameter Ta is set to 300 seconds. Based on the commit timestamp ti of each change record, all change records with the same change batch number within the time interval [ti, ti+Ta] are aggregated into a change transaction Txj. The change transaction contains a (key, ov, nv), where key represents the complete path of the configuration item, ov represents the old value, and nv represents the new value, so as to obtain an ordered sequence of change transactions.

[0012] For each numeric configuration item, collect the set of new values ​​in all its historical change transactions, and formulate a binning strategy for the set. Divide the elements in the set into three levels: low, medium, and high. Sort the elements in the set in ascending order, take the 33rd percentile as the low / medium level boundary, and take the 66th percentile as the medium / high level boundary.

[0013] For Boolean or enumerated configuration items, use the direct mapping method to symbolize them;

[0014] For each change item in the change transaction Txj, record its change direction to generate a change symbol;

[0015] The above process transforms the change transaction into a set of change symbols.

[0016] Furthermore, an improved temporal association rule mining algorithm is adopted to mine the change symbol set corresponding to the configuration change sequence. The minimum support is set to 0.005, the minimum confidence is set to 0.7, and the maximum time interval is 3 change transactions. The configuration change sequence is scanned to find all frequent 1-itemsets that satisfy the condition that the change support is greater than or equal to the minimum support. Candidate 2-item sequence patterns are generated through joining and pruning, and their change support and join confidence are calculated. The above process is iterated to generate longer sequence patterns until no new frequent itemsets are generated. Finally, all sequence patterns that satisfy the condition that the change support is greater than or equal to the minimum support and the join confidence is greater than or equal to the minimum confidence are output.

[0017] Furthermore, when a user wants to change the configuration item K from the current value Vc to the target value Vt, the tuple (Kt, Vc→Vt) is used as the query key to quickly find all sequence patterns in the configuration evolution pattern library index that precisely match the change symbol and have the same change direction.

[0018] Collect a set of changes to subsequent connections for all matching sequence patterns. For each change that appears in the collected data, calculate its comprehensive recommendation weight, which is the product of its corresponding connection confidence and relevance lift. Sort the changes in descending order by comprehensive recommendation weight and take the top 5 changes to generate a change recommendation list.

[0019] Furthermore, after the user selects some or all of the recommended items in the change recommendation list to form a set of changes to be executed, the system traverses all sequence patterns in the configuration evolution pattern library that are not directly related to the set of changes to be executed. For each such sequence pattern, the system checks each change item in the set of its subsequent connected change items. When an item with the same change symbol but a different change direction is found, the system alerts the user that the current change operation conflicts with historical experience.

[0020] Furthermore, for each configuration item K, the distribution of all different value symbols in the past 90 days is calculated to determine the information entropy of the configuration item. The expression for the information entropy is: U(K)=-Σ(px×log2px), where U(K) represents the information entropy of configuration item K and px represents the frequency of occurrence of the value symbol x.

[0021] For each configuration item K, the number of changes C(K) in the past 90 days is counted to analyze the stability weight of the configuration item. The expression for the stability weight is: Ws(K)=1 / [1+ln(1+C(K))], where Ws(K) represents the stability weight of configuration item K.

[0022] The system configuration entropy is analyzed based on the information entropy and stability weight of the configuration items. The expression of the system configuration entropy is: Hc=(Σ(U(K)×Ws(K))) / ΣWs(K), where Hc represents the system configuration entropy.

[0023] Furthermore, the ratio of the number of change items in the set of changes to the average number of change items in historical change transactions is used as the change size factor;

[0024] For each change item in the set of changes to be executed, the change symbol that appears most frequently in the configuration evolution pattern library corresponding to that change item is taken as the mainstream symbol. When the change symbol of the current analysis change item is the same as the mainstream symbol, the mainstream difference degree is set to 0; otherwise, the mainstream difference degree is set to 1. The ratio of the sum of the mainstream differences degree of all change items in the set of changes to be executed to the number of change items in the set of changes to be executed is taken as the change deviation factor.

[0025] The change set to be executed is matched with the sequence patterns in the configuration evolution pattern library. If it completely matches a sequence pattern in the configuration evolution pattern library with a connection confidence greater than 0.8, the pattern compliance factor is set to -0.5. If it partially matches a sequence pattern in the configuration evolution pattern library, the pattern compliance factor is set to 0. If it conflicts with a sequence pattern in the configuration evolution pattern library, the pattern compliance factor is set to 0.5.

[0026] Weights are assigned to the change size factor, change deviation factor, and pattern compliance factor to synthesize the expected entropy increase, where the expected entropy increase = 0.4 × change size factor + 0.4 × change deviation factor + 0.2 × pattern compliance factor.

[0027] Furthermore, the 30th percentile of the expected entropy increase of all successful historical change transactions is taken as the first risk threshold, the 80th percentile of the expected entropy increase of all successful historical change transactions is taken as the second risk threshold, and the 40th percentile of the system configuration entropy in a stable and healthy state is taken as the configuration entropy baseline. The ratio of the current system configuration entropy to the configuration entropy baseline is taken as the entropy state coefficient. The first and second risk thresholds are dynamically adjusted based on the entropy state coefficient. When the entropy state coefficient is in the range of [0.7, 1.5], the ratio of the first risk threshold to the entropy state coefficient is taken as the adjusted first risk threshold, and the ratio of the second risk threshold to the entropy state coefficient is taken as the adjusted second risk threshold.

[0028] The expected entropy increase is compared with the first risk threshold and the second risk threshold to classify the change risk level. When the expected entropy increase is less than or equal to the first risk threshold, the change risk level is set to low risk. When the expected entropy increases to the first risk threshold but is less than or equal to the second risk threshold, the change risk level is set to medium risk. When the expected entropy increases to the second risk threshold, the change risk level is set to high risk.

[0029] On the other hand, the present invention also provides a software configuration development management device based on artificial intelligence, comprising:

[0030] The extraction and processing module is used to extract the configuration change sequence from the database and perform feature processing.

[0031] The mining and analysis module is used to mine sequence patterns of configuration change sequences in order to build a configuration evolution pattern library;

[0032] The change recommendation module is used for intelligent change recommendations and conflict prediction based on sequence patterns;

[0033] The risk analysis module is used to determine the system configuration entropy based on the configuration evolution pattern library and to classify the change risk level.

[0034] On the other hand, the present invention also provides a storage medium storing instructions that, when run on a computer, cause the computer to execute the artificial intelligence-based software configuration development and management method described above.

[0035] The beneficial effects of this invention are as follows: By integrating feature extraction, sequence pattern mining, intelligent recommendation, and risk prediction mechanisms for configuration change sequences, intelligent understanding and proactive management of software configuration evolution are achieved. This method can automatically learn the association rules and evolutionary patterns between configuration items from historical changes, forming a queryable and interpretable configuration evolution pattern library, thereby providing accurate change recommendations and conflict warnings when changes are initiated. Simultaneously, by introducing configuration entropy and multi-dimensional risk factors, the system can quantitatively assess the stability of the configuration structure and the potential risks of change operations, and implement differentiated approval and control processes based on risk levels, significantly improving the reliability, automation level, and scientific nature of the change process, and reducing the risk of system failures caused by configuration errors. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a flowchart of the software configuration development and management method based on artificial intelligence in this embodiment.

[0038] Figure 2 This is a flowchart illustrating the modified intelligent recommendation and conflict prediction method in this embodiment.

[0039] Figure 3 This is a flowchart illustrating the method for classifying risk levels in this embodiment.

[0040] Figure 4 This is a schematic diagram of the structure of the AI-based software configuration development and management device in this embodiment. Detailed Implementation

[0041] The following detailed description, in conjunction with the accompanying drawings and specific embodiments, provides a further detailed explanation of the artificial intelligence-based software configuration development management method, apparatus, and storage medium disclosed in this invention. It should be noted that the technical features or combinations of technical features described in the following embodiments should not be considered isolated; they can be combined to achieve better technical effects. In the accompanying drawings of the following embodiments, the same reference numerals in each drawing represent the same features or components, which can be applied to different embodiments. Therefore, once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0042] It should be noted that the structures, proportions, sizes, etc., illustrated in the accompanying drawings are merely for illustrative purposes and to aid those skilled in the art in understanding and reading the invention. They are not intended to limit the conditions under which the invention can be implemented. Any modifications to the structure, changes in proportions, or adjustments to size, provided they do not affect the effectiveness or purpose of the invention, should fall within the scope of the technical content disclosed in the invention. The scope of the preferred embodiments of the present invention includes other implementations, wherein functions may be performed not in the order stated or discussed, including substantially simultaneously or in reverse order, depending on the functions involved. This should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0043] Techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and apparatus should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0044] In the description of the embodiments of this application, " / " means "or", and "and / or" is used to describe the relationship between related objects, indicating that there can be three relationships. For example, "A and / or B" means: A and B exist alone, B exists alone, and A and B exist simultaneously. In the description of the embodiments of this application, "multiple" refers to two or more embodiments.

[0045] Please see Figure 1 As shown, this is an artificial intelligence-based software configuration development and management method in this embodiment, including:

[0046] Step S1: Extract the configuration change sequence from the database and perform feature processing. The database is a configuration management database. The extracted fields include change ID, submission timestamp, operator, target service identifier, full path of configuration item, old value, new value, and change batch number. Generate a configuration change sequence based on the extracted content and set the time window parameter to 365 days. The time window parameter balances pattern coverage and data timeliness. A time window shorter than 180 days may result in insufficient data and failure to discover low-frequency but important stable patterns; a time window longer than 2 years may contain too many outdated change patterns that are inconsistent with the current system architecture. The time window parameter is an adjustable parameter and can be adjusted within the range of [180, 540] days according to the system change frequency during implementation.

[0047] Specifically, in step S1 of this embodiment, data cleaning is performed on the configuration change sequence. The data cleaning rules include filtering rollback operations, handling null values, and normalizing service identifiers.

[0048] The filtering rollback operation is as follows: records marked as old values ​​are the same as the current values ​​in the database, and records whose new values ​​are the same as earlier historical values ​​are rolled back and removed;

[0049] The handling of null values ​​is as follows: Changes to either the old or new value being NULL are uniformly represented by the same special symbol, such as... <null> ， <blank>etc., so as to regard it as a valid state change;

[0050] The service identifier normalization is to map and normalize the same logical service identifier in different environments, such as dev, test, prod, etc., to ensure that pattern mining is not interrupted due to environmental differences.

[0051] Specifically, in step S1 of this embodiment, the aggregation time window parameter Ta is set to 300 seconds. Based on the commit timestamp ti of each change record, all change records with the same change batch number within the time interval [ti, ti+Ta] are aggregated into a change transaction Txj. The change transaction contains a (key, ov, nv), where key represents the complete path of the configuration item, ov represents the old value, and nv represents the new value, to obtain an ordered sequence of change transactions. The change transaction sequence is [Tx1, Tx2, ..., TxM].

[0052] Specifically, in step S1 of this embodiment, for each numerical configuration item, a set of new values ​​in all its historical change transactions is collected, and a binning strategy for the set is formulated. The elements in the set are divided into three levels: low, medium, and high. The elements in the set are sorted in ascending order, and the 33rd percentile is taken as the low / medium level boundary, and the 66th percentile is taken as the medium / high level boundary.

[0053] For Boolean or enumeration type configuration items, use the direct mapping method to symbolize them, such as directly mapping the Boolean type true to ENABLED, directly mapping false to DISABLED, and directly using the enumeration type cache.type's [redis, memcached, local] as symbols, etc.

[0054] For each change item in the change transaction Txj, record its change direction to generate a change symbol. For example, if the old value symbol is A and the new value symbol is B and A≠B, then the generated change symbol is (K, A→B).

[0055] The above process transforms the change transaction into a set of change symbols.

[0056] Specifically, in step S1 of this embodiment, structured change records are extracted from the configuration management database, and the original change records are transformed into a standardized sequence of change symbols based on methods such as time window aggregation, data type binning, and symbolic mapping. This process effectively eliminates interference from noisy data, unifies service identifiers under different environments, and ensures the data quality and consistency for subsequent pattern mining. Through dynamic binning of numerical configuration items and symbolic representation of change directions, this step transforms continuous and discrete configuration values ​​into semantically meaningful symbolic states, providing a unified and computable feature representation for subsequent time-series correlation analysis and laying a reliable data foundation for pattern mining.

[0057] Please continue reading Figure 1 As shown, the AI-based software configuration development management method further includes:

[0058] Step S2: Mine the sequence patterns of configuration change sequences to build a configuration evolution pattern library.

[0059] Specifically, in step S2 of this embodiment, an improved temporal association rule mining algorithm is used to mine the set of change symbols corresponding to the configuration change sequence. The minimum support is set to 0.005, the minimum confidence is set to 0.7, and the maximum time interval is 3 change transactions. The configuration change sequence is scanned to find all frequent 1-itemsets that satisfy the condition that the change support is greater than or equal to the minimum support. Candidate 2-item sequence patterns are generated through joining and pruning, and their change support and join confidence are calculated. The above process is iterated to generate longer sequence patterns until no new frequent itemsets are generated. Finally, all sequence patterns that satisfy the condition that the change support is greater than or equal to the minimum support and the join confidence is greater than or equal to the minimum confidence are output. The 1-itemset refers to a single change symbol. The change support is the ratio of the number of times the same change item appears in historical change transactions to the total number of historical change transactions. The frequent itemset is the set of change items whose change support is greater than or equal to the minimum support. The connection confidence is the probability that when a change occurs, subsequent connected change items will follow. The sequence pattern is a triple (Tr, Co, St), where Tr represents the change symbol, Co represents the set of subsequent connected change items, and St represents the pattern statistics. The pattern statistics include change support, connection confidence, and correlation lift. The correlation lift = P(Co|Tr) / P(Co). A correlation lift greater than 1 indicates a positive correlation, and the larger the value, the stronger the correlation.

[0060] Specifically, in step S2 of this embodiment, each of the output sequence patterns is encapsulated as a pattern object and stored using a relational database or document database to build a configuration evolution pattern library. A composite index is built on the triples to ensure that real-time queries can respond in milliseconds.

[0061] Specifically, in step S2 of this embodiment, an improved temporal association rule mining algorithm is used to automatically discover frequently occurring configuration change patterns with temporal dependencies on the characteristic change symbol sequence. By setting parameters such as minimum support, confidence level, and maximum time interval, the algorithm can effectively identify recurring, high-confidence configuration linkage or sequential change patterns in history. The constructed configuration evolution pattern library encapsulates these patterns and their statistical indicators in a structured form and establishes an efficient index, thereby transforming implicit operational experience into an explicit, queryable knowledge base. This provides a core basis for real-time intelligent recommendation and risk analysis, significantly improving the interpretability and automation level of configuration management.

[0062] Please continue reading Figure 1 As shown, the AI-based software configuration development management method further includes:

[0063] Step S3: Perform intelligent change recommendations and conflict prediction based on sequence patterns.

[0064] Please see Figure 2 As shown, it is a method for changing intelligent recommendations and conflict prediction, including:

[0065] Step S31: Generate a change recommendation list based on the sequence pattern.

[0066] Specifically, in step S31 of this embodiment, when the user wants to change the configuration item K from the current value Vc to the target value Vt, the tuple (Kt, Vc→Vt) is used as the query key to quickly find all sequence patterns in the configuration evolution pattern library index that precisely match the change symbol and have the same change direction.

[0067] Collect a set of changes to subsequent connections for all matching sequence patterns. For each change that appears in the collected data, calculate its comprehensive recommendation weight, which is the product of its corresponding connection confidence and relevance lift. Sort the changes in descending order by comprehensive recommendation weight and take the top 5 changes to generate a change recommendation list.

[0068] Please continue reading Figure 2 As shown, the intelligent recommendation and conflict prediction method further includes:

[0069] Step S32: Collect user feedback on changes to the recommended list and perform conflict detection based on the configuration evolution pattern library.

[0070] Specifically, in step S32 of this embodiment, after the user selects some or all of the recommended items in the change recommendation list to form a set of changes to be executed, the system traverses all sequence patterns in the configuration evolution pattern library that are not directly related to the change symbols of the set of changes to be executed. For each such sequence pattern, the system checks each change item in the set of change items that are subsequently connected. When an item with the same change symbol but different change direction is found, the system alerts the user that the current change operation conflicts with historical experience.

[0071] Specifically, in step S3 of this embodiment, the configuration evolution pattern library is used to match similar historical change scenarios in real time when a user initiates a configuration change, automatically calculating and recommending highly relevant and high-confidence subsequent change items to form an intelligent recommendation list. This mechanism can help users discover potentially overlooked relevant configuration adjustments, improving the completeness and accuracy of changes. Simultaneously, by performing a conflict scan of the pattern library on the set of changes to be executed, this step can promptly identify potential conflicting operations that are inconsistent with historical successful patterns and proactively alert the user. This function transforms post-event conflict handling into pre-event risk prevention, effectively reducing the possibility of system anomalies caused by configuration inconsistencies or incompatibilities.

[0072] Please continue reading Figure 1 As shown, the AI-based software configuration development management method further includes:

[0073] Step S4: Determine the system configuration entropy based on the configuration evolution model library and classify the change risk level.

[0074] Please see Figure 3 As shown, this is a method for classifying risk levels, including:

[0075] Step S41: Analyze the system configuration entropy based on the configuration evolution pattern library.

[0076] Specifically, in step S41 of this embodiment, for each configuration item K, the distribution of all different value symbols in the past 90 days is calculated to determine the information entropy of the configuration item. The expression of the information entropy is: U(K)=-Σ(px×log2px), where U(K) represents the information entropy of configuration item K and px represents the frequency of occurrence of the value symbol x.

[0077] For each configuration item K, the number of changes C(K) in the past 90 days is counted to analyze the stability weight of the configuration item. The expression for the stability weight is: Ws(K)=1 / [1+ln(1+C(K))], where Ws(K) represents the stability weight of configuration item K.

[0078] The system configuration entropy is analyzed based on the information entropy and stability weight of the configuration items. The expression of the system configuration entropy is: Hc=(Σ(U(K)×Ws(K))) / ΣWs(K), where Hc represents the system configuration entropy.

[0079] Please continue reading Figure 3 As shown, the method for classifying change risk levels also includes:

[0080] Step S42: Based on the system configuration entropy analysis, predict the entropy increase.

[0081] Specifically, in step S42 of this embodiment, the ratio of the number of change items in the change set to the average number of change items in historical change transactions is used as the change scale factor.

[0082] For each change item in the set of changes to be executed, the change symbol that appears most frequently in the configuration evolution pattern library corresponding to that change item is taken as the mainstream symbol. When the change symbol of the current analysis change item is the same as the mainstream symbol, the mainstream difference degree is set to 0; otherwise, the mainstream difference degree is set to 1. The ratio of the sum of the mainstream differences degree of all change items in the set of changes to be executed to the number of change items in the set of changes to be executed is taken as the change deviation factor.

[0083] The change set to be executed is matched with the sequence patterns in the configuration evolution pattern library. If it completely matches a sequence pattern in the configuration evolution pattern library with a connection confidence greater than 0.8, the pattern compliance factor is set to -0.5. If it partially matches a sequence pattern in the configuration evolution pattern library, the pattern compliance factor is set to 0. If it conflicts with a sequence pattern in the configuration evolution pattern library, the pattern compliance factor is set to 0.5.

[0084] Weights are assigned to the change size factor, change deviation factor, and pattern compliance factor to synthesize the expected entropy increase, where the expected entropy increase = 0.4 × change size factor + 0.4 × change deviation factor + 0.2 × pattern compliance factor. The method for determining the conflict between the change set to be executed and the sequence patterns in the configuration evolution pattern library is the same as the conflict detection method in step S32 above.

[0085] Please continue reading Figure 3 As shown, the method for classifying change risk levels also includes:

[0086] Step S43: Classify the change risk level based on system configuration entropy and expected entropy increase.

[0087] Specifically, in step S43 of this embodiment, the 30th percentile of the expected entropy increase of all successfully changed historical change transactions is taken as the first risk threshold, the 80th percentile of the expected entropy increase of all successfully changed historical change transactions is taken as the second risk threshold, and the 40th percentile of the system configuration entropy in a stable and healthy state is taken as the configuration entropy baseline. The ratio of the current system configuration entropy to the configuration entropy baseline is taken as the entropy state coefficient. The first and second risk thresholds are dynamically adjusted based on the entropy state coefficient. When the entropy state coefficient is within [0.7, 1.5], the ratio of the first risk threshold to the entropy state coefficient is taken as the adjusted first risk threshold, and the ratio of the second risk threshold to the entropy state coefficient is taken as the adjusted second risk threshold. The stable and healthy state of the system refers to a preset system operation period, such as a period without major failures or a stable operation period, the purpose of which is to collect data from the system under normal operating conditions to analyze baseline data.

[0088] Specifically, in step S43 of this embodiment, the expected entropy increase is compared with the first risk threshold and the second risk threshold to classify the change risk level. When the expected entropy increase is less than or equal to the first risk threshold, the change risk level is set to low risk. When the expected entropy increases to the first risk threshold but is less than or equal to the second risk threshold, the change risk level is set to medium risk. When the expected entropy increases to the second risk threshold, the change risk level is set to high risk.

[0089] Specifically, in step S43 of this embodiment, regarding the classification of change risk levels, for low risk, the process is automated, and the system automatically submits the change to the configuration center; for medium risk, the process is interrupted, triggering an approval chain, and the system sends the change details, intelligent recommendations, conflict prediction results, and entropy increase calculation details to the person in charge of the team to which the service belongs or the on-duty maintenance personnel. After the approver confirms, the process continues; for high risk, the process is forcibly blocked, and the system requires the change initiator to fill out a "High-Risk Change Request Form," which must include: a description of the business urgency, a detailed technical solution, a monitoring and rollback plan, and the upstream and downstream service teams that have been notified. The request form must be approved by a higher-level architect or department head.

[0090] Specifically, in step S4 of this embodiment, information entropy theory is innovatively introduced to quantify the disorder of the system configuration structure. Stability weights are calculated based on the change frequency of configuration items, ultimately yielding the system configuration entropy, which is used to comprehensively assess the stability of the configuration baseline. Furthermore, by comprehensively considering the change scale, the degree of deviation from the mainstream pattern, and the matching with historical patterns, the projected entropy increase is synthesized to quantify the disorder risk that a single change may introduce. Based on the entropy increase distribution of historical successful changes and the current system entropy state, risk thresholds are dynamically set to achieve low, medium, and high risk classification. This step upgrades risk control from qualitative judgment to quantitative assessment and triggers differentiated process controls based on different risk levels, achieving refined and adaptive management of change risks and greatly enhancing the safety and controllability of system changes.

[0091] Please see Figure 4 As shown, this is an artificial intelligence-based software configuration development and management device in this embodiment, including:

[0092] The extraction and processing module is used to extract the configuration change sequence from the database and perform feature processing.

[0093] The mining and analysis module is used to mine sequence patterns of configuration change sequences in order to build a configuration evolution pattern library;

[0094] The change recommendation module is used for intelligent change recommendations and conflict prediction based on sequence patterns;

[0095] The risk analysis module is used to determine the system configuration entropy based on the configuration evolution pattern library and to classify the change risk level.

[0096] This application also provides a computer-readable storage medium storing instructions that, when run on a computer, cause the computer to execute the AI-based software configuration development and management method as described in the above method embodiments.

[0097] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable programs, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable programs, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0098] In the above description, the disclosure of this invention is not intended to limit itself to these aspects. Rather, within the scope of the objectives of this disclosure, components can be selectively and operationally combined in any number. Furthermore, terms such as "comprising," "encompassing," and "having" should be interpreted by default as inclusive or open-ended, rather than exclusive or closed, unless explicitly defined as such. All technical, scientific, or other terms are to be understood by those skilled in the art, unless defined as such. Public terms found in dictionaries should not be interpreted in a too idealistic or impractical manner in the context of the relevant technical documents, unless explicitly defined as such in this disclosure. Any modifications or alterations made by those skilled in the art based on the foregoing disclosure are within the scope of the claims.< / blank> < / null>

Claims

1. A software configuration development management method based on artificial intelligence, characterized in that, include: Extract the configuration change sequence from the database and perform feature processing; Discover sequence patterns in configuration change sequences to build a configuration evolution pattern library; Intelligent change recommendation and conflict prediction based on sequence patterns; The system configuration entropy is determined based on the configuration evolution pattern library, and the change risk level is classified.

2. The software configuration development management method based on artificial intelligence according to claim 1, characterized in that, Set the aggregation time window parameter Ta=300 seconds. Based on the commit timestamp ti of each change record, aggregate all change records with the same change batch number within the time interval [ti,ti+Ta] into a change transaction Txj. The change transaction contains a (key,ov,nv), where key represents the complete path of the configuration item, ov represents the old value, and nv represents the new value, so as to obtain an ordered sequence of change transactions. For each numeric configuration item, collect the set of new values ​​in all its historical change transactions, and formulate a binning strategy for the set. Divide the elements in the set into three levels: low, medium, and high. Sort the elements in the set in ascending order, take the 33rd percentile as the low / medium level boundary, and take the 66th percentile as the medium / high level boundary. For Boolean or enumerated configuration items, use the direct mapping method to symbolize them; For each change item in the change transaction Txj, record its change direction to generate a change symbol; The above process transforms the change transaction into a set of change symbols.

3. The software configuration development management method based on artificial intelligence according to claim 2, characterized in that, An improved temporal association rule mining algorithm is adopted to mine the change symbol set corresponding to the configuration change sequence. The minimum support is set to 0.005, the minimum confidence is set to 0.7, and the maximum time interval is 3 change transactions. The configuration change sequence is scanned to find all frequent one-itemsets that satisfy the condition that the change support is greater than or equal to the minimum support. Candidate two-item sequence patterns are generated through joining and pruning, and their change support and join confidence are calculated. The above process is iterated to generate longer sequence patterns until no new frequent itemsets are generated. Finally, all sequence patterns that satisfy the condition that the change support is greater than or equal to the minimum support and the join confidence is greater than or equal to the minimum confidence are output.

4. The software configuration development management method based on artificial intelligence according to claim 3, characterized in that, When a user wants to change the configuration item K from the current value Vc to the target value Vt, the tuple (Kt, Vc→Vt) is used as the query key to quickly find all sequence patterns in the configuration evolution pattern library index that precisely match the change symbol and have the same change direction. Collect a set of changes to subsequent connections for all matching sequence patterns. For each change that appears in the collected data, calculate its comprehensive recommendation weight, which is the product of its corresponding connection confidence and relevance lift. Sort the changes in descending order by comprehensive recommendation weight and take the top 5 changes to generate a change recommendation list.

5. The software configuration development management method based on artificial intelligence according to claim 4, characterized in that, After the user selects some or all of the recommended items in the change recommendation list to form a set of changes to be executed, the system traverses all sequence patterns in the configuration evolution pattern library that are not directly related to the set of changes to be executed. For each such sequence pattern, the system checks each change item in the set of its subsequent connected change items. When an item with the same change symbol but different change direction is found, the system alerts the user that the current change operation conflicts with historical experience.

6. The software configuration development management method based on artificial intelligence according to claim 5, characterized in that, For each configuration item K, calculate the distribution of all different value symbols in the past 90 days to determine the information entropy of the configuration item. The expression for the information entropy is: U(K)=-Σ(px×log2px), where U(K) represents the information entropy of configuration item K and px represents the frequency of occurrence of the value symbol x. For each configuration item K, the number of changes C(K) in the past 90 days is counted to analyze the stability weight of the configuration item. The expression for the stability weight is: Ws(K)=1 / [1+ln(1+C(K))], where Ws(K) represents the stability weight of configuration item K. The system configuration entropy is analyzed based on the information entropy and stability weight of the configuration items. The expression of the system configuration entropy is: Hc=(Σ(U(K)×Ws(K))) / ΣWs(K), where Hc represents the system configuration entropy.

7. The software configuration development management method based on artificial intelligence according to claim 6, characterized in that, The ratio of the number of change items in the pending change set to the average number of change items in historical change transactions is used as the change size factor; For each change item in the set of changes to be executed, the change symbol that appears most frequently in the configuration evolution pattern library corresponding to that change item is taken as the mainstream symbol. When the change symbol of the current analysis change item is the same as the mainstream symbol, the mainstream difference degree is set to 0; otherwise, the mainstream difference degree is set to 1. The ratio of the sum of the mainstream differences degree of all change items in the set of changes to be executed to the number of change items in the set of changes to be executed is taken as the change deviation factor. The change set to be executed is matched with the sequence patterns in the configuration evolution pattern library. If it completely matches a sequence pattern in the configuration evolution pattern library with a connection confidence greater than 0.8, the pattern compliance factor is set to -0.

5. If it partially matches a sequence pattern in the configuration evolution pattern library, the pattern compliance factor is set to 0. If it conflicts with a sequence pattern in the configuration evolution pattern library, the pattern compliance factor is set to 0.

5. Weights are assigned to the change size factor, change deviation factor, and pattern compliance factor to synthesize the expected entropy increase, where the expected entropy increase = 0.4 × change size factor + 0.4 × change deviation factor + 0.2 × pattern compliance factor.

8. The software configuration development management method based on artificial intelligence according to claim 7, characterized in that, The first risk threshold is set as the 30th percentile of the expected entropy increase of all successful historical change transactions, the second risk threshold is set as the 80th percentile of the expected entropy increase of all successful historical change transactions, and the third risk threshold is set as the 40th percentile of the system configuration entropy in a stable and healthy state. The ratio of the current system configuration entropy to the configuration entropy baseline is set as the entropy state coefficient. The first and second risk thresholds are dynamically adjusted based on the entropy state coefficient. When the entropy state coefficient is in the range of [0.7, 1.5], the ratio of the first risk threshold to the entropy state coefficient is set as the adjusted first risk threshold, and the ratio of the second risk threshold to the entropy state coefficient is set as the adjusted second risk threshold. The expected entropy increase is compared with the first risk threshold and the second risk threshold to classify the change risk level. When the expected entropy increase is less than or equal to the first risk threshold, the change risk level is set to low risk. When the expected entropy increases to the first risk threshold but is less than or equal to the second risk threshold, the change risk level is set to medium risk. When the expected entropy increases to the second risk threshold, the change risk level is set to high risk.

9. An artificial intelligence-based software configuration development management device, applied to the artificial intelligence-based software configuration development management method as described in any one of claims 1-8, characterized in that, include: The extraction and processing module is used to extract the configuration change sequence from the database and perform feature processing. The mining and analysis module is used to mine sequence patterns of configuration change sequences in order to build a configuration evolution pattern library; The change recommendation module is used for intelligent change recommendations and conflict prediction based on sequence patterns; The risk analysis module is used to determine the system configuration entropy based on the configuration evolution pattern library and to classify the change risk level.

10. A storage medium, characterized in that, It stores instructions that, when run on a computer, cause the computer to perform the AI-based software configuration development management method as described in any one of claims 1-8.

Images