Method for constructing an internet-based enterprise service application system

Abstract

The present application relates to the technical field of enterprise digitization service, in particular to a construction method of enterprise service application system based on Internet, which firstly constructs a minimum atomization system base, completes the standardized atomic encapsulation of the minimum granularity of business capability and the construction of full-dynamic microkernel architecture; secondly, the full-dimension operation carbon footprint of the user is collected without sensing and semantic analysis is completed, which is converted into structured business intent data; then a unified demand decision hub is built, two types of demands of function growth expansion and optimization contraction are synchronously identified and priority quantization sorting is completed; finally, double-line self-adaptive optimization is executed in parallel based on the atomization base, and all optimizations are completed in the way of hot deployment. The present application realizes the dynamic adaptation of enterprise service system and business development, and solves the problem of rigidity of traditional system architecture, disconnection between iteration and business demand.

Description

Technical Field

[0001] This invention relates to the field of enterprise digital service technology, and more specifically, to a method for constructing an Internet-based enterprise service application system. Background Technology

[0002] With the comprehensive advancement of enterprise digital transformation, internet-based enterprise service application systems have become the core carriers for enterprises to achieve business digitalization, process standardization, and refined management. Their architecture design and iteration capabilities directly determine the implementation effect and adaptation efficiency of enterprise digital transformation.

[0003] Existing enterprise service systems mostly adopt a stable, pre-defined architecture, developing fixed functions and designing the architecture based on business needs in the early stages of transformation. They have never undergone adaptive evolution design at the architectural level to address the phased and dynamic characteristics of enterprise digital transformation. Adding, deleting, and optimizing system functions all require manual intervention and service shutdowns for releases, making it difficult to adapt to the rapid iteration needs of enterprise businesses. This easily leads to problems such as rigid architecture, functional redundancy, and business gaps. Furthermore, existing technologies rely heavily on manual surveys and business feedback to identify system iteration needs. This only enables correlation analysis between system governance and business requirements, failing to quantify the causal relationship between business intent and needs based on actual user behavior. It is difficult to deeply bind system governance actions with enterprise business goals, easily resulting in a disconnect between system iteration and actual business needs.

[0004] To address the shortcomings of the existing technologies, this invention proposes a method for constructing an Internet-based enterprise service application system. Summary of the Invention

[0005] In view of the shortcomings of existing technologies, the purpose of this invention is to provide a method for constructing an Internet-based enterprise service application system.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] The method for building an internet-based enterprise service application system includes the following steps:

[0008] S1. Construct the foundation of the minimum atomic system, conduct architectural analysis and business capability decomposition of the target enterprise's full business domain, complete the standardized encapsulation of the minimum granularity atomic capabilities, form an atomic capability library that can be independently orchestrated, build a fully dynamic microkernel architecture that retains only basic management and control capabilities and supports hot-swappable access to business functions, and deploy the initial minimum feasible baseline system based on the atomic capabilities necessary for the enterprise's core business.

[0009] S2. Collect and semantically analyze user operation carbon footprints. Collect user operation behavior data and corresponding context information in all dimensions without being aware of the user. Transform the discrete raw operation logs into structured business intent data. At the same time, generate standardized tags for users based on the analysis results and complete user group segmentation. Output structured data that can be used for demand identification.

[0010] S3. Establish a unified demand decision center, synchronously identify two types of demands—functional growth and expansion, and optimization and contraction—based on semantic parsing results, construct a multi-dimensional priority scoring model to uniformly quantify and sort all demands, and automatically generate corresponding structured demand specifications.

[0011] S4. Based on the aforementioned atomized system base, dual-path adaptive optimization is performed in parallel, involving demand-driven functional growth and expansion as well as behavior-driven optimization and contraction. All optimization actions are completed using a hot deployment method.

[0012] Furthermore, in step S1, the atomic capability is decomposed to the smallest indivisible business action node. After verifying the business boundary of the node, the corresponding execution logic and data interaction rules are matched to complete the initial encapsulation of the atomic capability. The encapsulated atomic capability has independent containerized deployment, independent start and stop management, standardized input and output interfaces, clear resource consumption boundaries, and full-link traceability.

[0013] Furthermore, the fully dynamic microkernel architecture predefines the standardized mapping relationship between interface element types and atomic capability interfaces. It can retrieve the corresponding list of accessible functions according to user permission configuration, group by business domain, and complete the layout and display of interface elements according to the function usage priority. The addition, deletion and modification of functions trigger the interface to automatically adapt synchronously.

[0014] Furthermore, in step S2, the collected operational behavior data includes explicit business operation data, implicit behavioral preference data, cross-system off-site operation data, and operation context information. After all collected data is standardized in format, it is stored in the system log library as a data source for semantic parsing.

[0015] Furthermore, in step S2, semantic parsing first needs to construct a standardized mapping relationship between operation sequences and business intentions, complete the matching and transformation of discrete operation sequences into business intentions, then perform cluster statistics on the matched business intentions, and at the same time calculate the efficiency coefficient of the corresponding operation path and mark inefficient detour operations.

[0016] Furthermore, in step S3, the identified functional growth and expansion requirements include high-frequency atomic operation serialization requirements, business capability coverage gap requirements, and best practice replication requirements; the identified optimization and contraction requirements include low-frequency redundant function reduction requirements, duplicate and overlapping function merging requirements, and inefficient process optimization requirements.

[0017] Furthermore, the identification of repeated and overlapping functions uses the function call sequence within a single user session as a transaction set, iteratively mines frequent itemsets in the transaction set and generates association rules between functions, calculates function redundancy based on the support and confidence of the bidirectional association rules, and completes the determination of repeated and overlapping functions.

[0018] Furthermore, the execution of the functional growth and expansion path, in response to the high-frequency atomic operation sequence combination requirements, first assesses the data flow feasibility, transaction consistency, and exception handling compatibility of the atomic capabilities within the sequence. After the assessment is passed, the aggregation service code is automatically generated to complete the dynamic hot deployment and full-link traceability management of the function.

[0019] Furthermore, the execution of the optimized shrinkage route involves maintaining a dedicated forgetting curve for each low-frequency redundant function and quantifying the degree of forgetting. Based on the function usage interval and the degree of forgetting, a multi-stage progressive degradation action is set, and corresponding function recovery mechanisms and hierarchical offline management rules are configured.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] 1. This invention achieves a quantitative correlation between system governance actions and business objectives at the causal level by collecting full user operation carbon footprint data, semantic parsing, and establishing a unified demand decision-making center. This solves the problem that existing technologies can only perform correlation analysis and system iteration is disconnected from business needs. The method collects user operation behavior data across all dimensions without any awareness, transforms discrete operation logs into structured business intent data, synchronously identifies two types of core needs based on data from the same source, and completes the quantitative ranking of needs through a multi-dimensional priority scoring model. This ensures that all system optimization actions are driven by real business needs, achieving a deep binding between governance actions and business objectives.

[0022] 2. This invention employs a dual-path adaptive optimization solution that combines demand-driven functional growth and expansion with behavior-driven optimization and contraction. Based on a minimal atomized system foundation and a fully dynamic hot-swappable microkernel architecture, it achieves orderly functional growth through three types of actions: high-frequency atomic operation serialization, business gap filling, and best practice replication. Simultaneously, it achieves healthy system contraction through three types of actions: gradual decommissioning via the forgetting curve, merging redundant functions, and refactoring inefficient processes. All optimizations are completed using hot deployment. This solution fundamentally overturns the pre-defined logic of traditional steady-state governance architectures, achieving architecture-level adaptive evolution tailored to the phased characteristics of enterprise digital transformation. It achieves dynamic adaptation of the governance system to the transformation pace, resolving the core pain points of rigid traditional system architectures, functional expansion, and coexisting business gaps. Attached Figure Description

[0023] Figure 1A flowchart illustrating the construction method of an Internet-based enterprise service application system;

[0024] Figure 2 This is a flowchart illustrating the implementation of step S3 of the present invention: unifying the demand decision-making center and generating demand quantification.

[0025] Figure 3 This is a flowchart illustrating the implementation of the adaptive optimization execution of the dual-path parallel process in step S4 of the present invention. Detailed Implementation

[0026] Example, refer to Figure 1 The method for constructing an Internet-based enterprise service application system in this embodiment specifically includes the following steps:

[0027] S1, Construction of the Minimal Atomization System Base.

[0028] The standardized atomic capability library and fully dynamic hot-swappable kernel output in this step serve as the execution carrier for all subsequent data acquisition, requirement identification, and adaptive optimization actions, forming a data flow and control flow relationship between the foundational support and upper-level execution of subsequent steps.

[0029] S11. Decomposition and Standardized Encapsulation of Business Domain Atomic Capabilities:

[0030] First, conduct an architecture analysis of the target enterprise's entire business domains to identify the core business domains and supporting business domains. Decompose the business capabilities of each business domain into the smallest granularity atomic capabilities. Each atomic capability strictly follows a unified encapsulation specification to ensure complete decoupling and independent orchestration.

[0031] The specific steps for disassembling atomic capabilities are as follows:

[0032] The first step is to analyze all business processes in the business domain and break them down into the smallest indivisible business action nodes, with each node corresponding to only one complete and independent business operation.

[0033] The second step is to verify the business boundaries of each action node to ensure that there are no two or more business operations within the node that do not have direct data dependencies.

[0034] The third step is to match the corresponding execution logic and data interaction rules for each action node to ensure that a single node can independently complete the corresponding business operation without relying on the running logic of other nodes.

[0035] The fourth step is to assign a unique functional identifier to each node, completing the initial encapsulation of atomic capabilities;

[0036] The encapsulation specifications for atomic capabilities include: independent containerized deployment capabilities, operating environments that do not depend on other atomic capabilities; independent start-up and shutdown management capabilities, where individual enabling, disabling, and decommissioning operations do not affect the normal operation of other system functions; standardized input and output interfaces, unified input parameter, output parameter, and exception return format specifications, supporting on-demand combination and chaining of any atomic capabilities; clear resource consumption boundaries, quantifying CPU, memory, and storage resource consumption thresholds, supporting independent resource usage monitoring and management; and end-to-end traceability capabilities, enabling end-to-end traceability of all calls, combinations, and synthesis operations through a unique functional identifier.

[0037] S12, Fully Dynamic Hot-Swap Kernel Architecture Implementation:

[0038] The system kernel adopts a microkernel design, retaining only five basic capabilities: component registration, message communication, lifecycle management, access control, and dynamic interface assembly. All business functions are connected to the kernel in the form of atomic capability components.

[0039] The kernel's implementation logic includes: a dynamic component registration mechanism, supporting the dynamic registration, installation, upgrade, and uninstallation of atomic capability components during system runtime without requiring service shutdowns or kernel refactoring; a dynamic service orchestration mechanism, supporting the on-demand chaining, paralleling, and combination of multiple atomic capabilities based on standardized interfaces without modifying the atomic capability's own code; a dynamic interface assembly mechanism, where the system interface is automatically generated based on currently enabled atomic capabilities and combined functions, with the addition, decommissioning, and merging of functions triggering automatic interface adaptation without manual development adjustments; and a full lifecycle status management mechanism, monitoring the running status, usage rate, and resource consumption of all atomic capabilities and combined functions across the entire chain, synchronizing the collected status data to subsequent requirement analysis and optimization steps.

[0040] The dynamic interface assembly mechanism is implemented as follows: The system predefines a standardized mapping relationship between interface element types and atomic capability interfaces, with each atomic capability corresponding to a set of independently renderable basic interface elements; the system retrieves a list of atomic capabilities and combination functions accessible to the currently logged-in user based on their permission configuration; the system groups the function list by business domain and automatically determines the layout order and display hierarchy of interface elements based on the frequency and priority of function usage. Priority is calculated using a priority score, with the formula as follows: The normalization process maps the values ​​to the range of 0 to 1, and the weight allocation rules are clearly defined. In this embodiment, the preferred weights are: 1.5 for the business domain function of the current session, 1.2 for the high-frequency function corresponding to the core user role, 1.0 for the general auxiliary function, and 0.8 for the low-frequency edge function. In addition, the layout and sorting rules are: functions with a weighted score ≥ 0.8 are fixed in the core area at the top of the first-level menu; functions with a score of 0.3 to 0.8 are arranged in descending order in the regular area of ​​the first-level menu; functions with a score < 0.3 are included in the second-level menu; and functions with a score < 0.1 are only displayed in the global search.

[0041] The addition, removal, and merging of features trigger the synchronous addition, deletion, and layout rearrangement of corresponding interface elements, without the need for manual modification of the front-end code;

[0042] S13. Initial Baseline System Deployment:

[0043] Based on the fully encapsulated atomic capability library and hot-swappable kernel, an initial minimum feasible baseline system is built. Only the atomic capabilities essential to the core business of the enterprise are deployed, without pre-setting any unnecessary functions, thus avoiding over-design of the initial system. Simultaneously, it provides a runnable and data-collectible basic operating environment for subsequent user operation data collection and dual-route optimization. The essential atomic capabilities must simultaneously meet the following three quantitative conditions to be included in the initial baseline system:

[0044] (1) Business coverage: Corresponding to the main revenue process of the enterprise, covering ≥80% of the core business departments;

[0045] (2) Rigidity: It is a necessary node in the business process, with no alternative execution path, and the call rate is 100% when the process is executed;

[0046] (3) Scope of impact: The disabling of the function will cause the corresponding main process to be completely interrupted, affecting ≥60% of the system's active users.

[0047] S2. User operation carbon footprint collection and semantic parsing.

[0048] It takes user operation data generated by the baseline system built by S1 and transforms it from discrete and meaningless operation logs into structured business requirements. The output semantic intent data and user profile data will serve as the sole input for the S3 requirement decision-making steps, realizing the objectivity and full quantification of requirement collection.

[0049] S21. Seamless full-scale carbon footprint data collection:

[0050] The operation carbon footprint referred to in this embodiment refers to the set of operation behavior footprint data of users in enterprise service application systems, including but not limited to: user identifier, session identifier, timestamp, page / function identifier, control identifier, operation type, input parameter summary, return result code and context information associated with the operation (such as business domain, role permissions, terminal type, network status). The operation carbon footprint is used to reconstruct the session, semantically parse and quantify the requirements of the user operation sequence.

[0051] Fine-grained tracking is used to collect user actions across all dimensions within the system and related business scenarios. The collection process is seamless for users and does not affect normal business operations. The collected data includes four categories: First, explicit business operation data, including core business actions such as clicks, input, submission, export, printing, approval, and querying, which is collected in full through front-end tracking and back-end service logs. Second, implicit behavioral preference data, including page dwell time, scrolling behavior, mouse trajectory, shortcut key usage, erroneous operations, and help document viewing frequency, is collected through front-end tracking. Third, cross-system off-site operation data, including Excel data processing and re-import, third-party system redirects, and offline business operations authorized by users, is collected through client-side proxies or browser plugins. Fourth, operation context information, including the current page URL, operation object ID, operation timestamp, user role, department, and operation terminal, is automatically associated and completed by the system. All collected data is standardized in format and stored in the system log library to provide a data source for subsequent semantic parsing.

[0052] S22. Semanticization of Operation Sequences and Intent Parsing:

[0053] The standardized raw operation logs are transformed into quantifiable and understandable structured business requirements. The specific execution steps are as follows:

[0054] The first step is to construct a standardized mapping relationship between operation sequences and business intentions, and to complete the matching and transformation of discrete operation sequences into business intentions. Each business intention corresponds to a unique combination of continuous operation sequences.

[0055] The second step is intent clustering and statistics. The DBSCAN clustering algorithm is used to cluster similar operation sequences into the same intent cluster, and the frequency of occurrence, average time, number of users covered, and cross-departmental distribution data of each intent cluster are statistically analyzed.

[0056] The complete execution steps of the DBSCAN clustering algorithm are as follows:

[0057] 1. Using a single complete operation sequence as a sample, the edit distance between the operation sequences is used as the distance metric between the samples. The edit distance is calculated by the minimum number of insertion, deletion, and replacement operations between two operation sequences.

[0058] 2. Dynamically adjust the neighborhood radius. The neighborhood radius is a preset proportion of the average length of all operation sequences in the current sample set. The proportion can be adjusted according to the average complexity of the operation sequences. The preferred default value is 0.4, and the value range is 0.3~0.5. The determination method is as follows: 0.5 for simple scenarios with an average operation sequence length ≤ 10 steps, 0.4 for medium scenarios with 10~30 steps, and 0.3 for complex scenarios with > 30 steps. When the proportion of noise points is > 15%, the proportion is reduced by 0.05, and when the number of clusters is < 5, the proportion is increased by 0.05.

[0059] 3. Determine the core point determination rules. The minimum neighborhood sample number of a core point is determined based on the total number of active users in the system. If the number of samples contained in the neighborhood of a sample is greater than or equal to the minimum neighborhood sample number, then that sample is a core point. The formula for calculating the minimum neighborhood sample number of a core point is as follows: The rule is that when the total number of active users is ≤1000, the value is fixed at 5; when the number of users is 1000~10000, the value is rounded up to 0.5%; and when the number of users is >10000, the upper limit is fixed at 50.

[0060] 4. Starting from any unvisited core point, traverse all density-reachable samples to form a cluster, completing the intent clustering of all operation sequences. Each cluster corresponds to a business intent.

[0061] The third step is to evaluate the efficiency of the operation path. This involves calculating the efficiency coefficient of the operation path for each intent using a quantitative formula to identify inefficient detours. The formula is as follows:

[0062] ;

[0063] In the formula, This is the operation path efficiency coefficient, with a value range of [value range missing]. ; The theoretically optimal number of operation steps corresponding to the business intent is determined by the minimum necessary number of nodes in the business process, and the value is a positive integer. The actual number of steps a user takes to complete this business intent, and the value is a positive integer; when When the efficiency coefficient is less than the preset threshold, the operation sequence is marked as a detour operation and included in the scope of subsequent optimization requirements. The threshold can be determined according to the enterprise's business characteristics and process complexity, with a preferred default value of 0.6. The determination method is as follows: 0.7 for simple processes with ≤3 optimal operation steps, 0.6 for medium processes with 4~10 steps, and 0.5 for complex processes with >10 steps.

[0064] S23. User profiling and user segmentation:

[0065] Based on the semantically parsed operational data, standardized tags are automatically generated for users, providing a basis for personalized optimization for different user groups. The generated tags include three categories: the first category is role tags, which infer the user's corresponding business role from the functional modules they access; the second category is proficiency tags, which assess the user's system operation proficiency level based on the user's operation speed, shortcut key usage rate, error rate, and number of times the help document is viewed; and the third category is preference tags, which statistically analyze the user's frequently used function ranking and interface layout preference data.

[0066] The specific logic for tag generation is as follows: Role tags are generated by statistically analyzing the business domains corresponding to the functional modules accessed by users, and matching them with a preset mapping relationship between business roles and functional domains. In this embodiment, preferably, the user's single business domain function access time accounts for ≥60%, and the role tag corresponding to that business domain is matched. Proficiency tags are generated by performing min-max normalization on four indicators: user operation speed, shortcut key usage rate, error rate, and number of times help documents are viewed, and then performing a weighted sum according to preset weights to obtain a proficiency score. Proficiency levels are generated based on the score range, and the weights can be adjusted according to business characteristics. The normalization process maps all indicator values ​​to the range of 0~1. The details are as follows:

[0067] The baseline weights for proficiency indicators are: operation speed 0.35, shortcut key usage rate 0.25, error rate 0.25, and help document viewing count 0.15. Proficiency levels are quantified based on a score between 0 and 1, as follows: Expert level (≥0.9), suitable for users highly proficient in system operation; Proficient level (0.7-0.89), suitable for users with moderately proficient operation; Average level (0.4-0.69), suitable for users with basic operational skills; Beginner level (0.2-0.39), suitable for users in the learning phase; Difficult level (<0.2), suitable for users with significant operational difficulties.

[0068] Preference tags are generated by sorting user function usage frequency and the distribution ratio of interface operation modes. After the tags are generated, users are divided into groups according to the combination of tags, which provides a basis for subsequent group demand identification and optimization.

[0069] S3, Unified Demand Decision-Making Center and Demand Quantification Generation.

[0070] like Figure 2 As shown, this step takes the semantic operation data and user group data output by S2, and simultaneously identifies two types of needs: functional growth and contraction. It completes the priority ranking of needs through a quantitative model, and solves the problem of demand conflict between the two routes from the source. The output structured demand specifications will be directly used as the input for the S4 dual-route execution steps, ensuring that all optimization actions are driven by real business value.

[0071] S31. Simultaneous identification of dual-dimensional requirements:

[0072] Based on the S2 analysis results, two types of core requirements are identified simultaneously, ensuring that the data sources for growth and contraction actions are from the same source, thus avoiding requirement conflicts from the outset.

[0073] The first category is functional growth and expansion needs, which are triggered by three scenarios: First, high-frequency atomic operation sequence needs, where users frequently execute fixed combinations of atomic capabilities, and the existing system cannot retain these combinations, so they need to be combined into new aggregated functions; second, business capability coverage gap needs, where users frequently complete business actions through cross-system off-site operations or bypass operations, and the existing atomic capabilities cannot cover these business needs, so new atomic capabilities need to be added; and third, best practice replication needs, where a high-frequency and efficient operation sequence of a certain department or role has enterprise-wide applicability and can be combined into a general function for promotion throughout the enterprise.

[0074] The second category is functional optimization and contraction requirements, which are triggered by three scenarios: First, low-frequency redundant functional requirements, where atomic capabilities or combination functions are not used for a long time and continuously occupy system resources, requiring gradual reduction and decommissioning; second, repetitive and overlapping functional requirements, where multiple atomic capabilities or combination functions have highly overlapping business scenarios and low usage rates, requiring merging and integration; and third, inefficient process optimization requirements, where user operation paths are circuitous and approval processes have serial bottlenecks, requiring compression, merging, and optimization.

[0075] The identification of duplicate and overlapping functions is implemented using the Apriori association rule algorithm. The complete execution steps are as follows:

[0076] 1. Construct the algorithm input dataset using the sequence of function calls within a single user session as the transaction set and the unique identifier of the function as the item;

[0077] 2. Iteratively mine frequent itemsets in the transaction set. The minimum support can be determined based on the total number of system functions and the scale of user activity. The support is the proportion of the number of transactions containing the corresponding itemset to the total number of transactions. When the average monthly total number of sessions is ≤5000, the minimum support is fixed at 0.02. When the average monthly total number of sessions is 5000~100000, the minimum support is calculated according to the formula. When the average monthly total number of sessions is >100000, the upper limit of the minimum support is fixed at 0.01.

[0078] 3. Based on the association rules between frequent itemset generation functions, the minimum confidence level can be determined according to the business scenario. The confidence level is the proportion of transactions that contain the preceding item and also contain the following item. In this embodiment, the preferred confidence level is 0.8 for core main business processes, 0.6 for edge auxiliary business processes, and 0.7 for general utility functions.

[0079] 4. Two-way association rule determination: If two functions simultaneously have both positive and negative association rules with high support and high confidence, then the two functions are determined to be redundant functions with highly overlapping business scenarios. The determination formula is as follows:

[0080] ;

[0081] In the formula, Scoring the redundancy of function a and function b, the higher the score, the greater the redundancy of the two functions; The support of the association rule from function a to function b; The support of the association rules from function b to function a; The confidence level of the association rule from function a to function b; For the confidence of the association rule from function b to function a, when When the redundancy exceeds the preset redundancy threshold, the two functions are determined to be duplicate and overlapping functions that need to be merged; the redundancy threshold is determined by the following rules: the baseline default value is 0.45, the threshold for core business functions is increased to 0.6, and the threshold for edge auxiliary functions is decreased to 0.3.

[0082] S32. Quantitative ranking of demand priorities:

[0083] A unified multi-dimensional priority scoring model is constructed to quantitatively score all growth and contraction requirements, and the requirements are ranked from highest to lowest score to determine their execution priority. The scoring formula is as follows:

[0084] ;

[0085] In the formula, The requirements are scored based on their priority; the higher the score, the higher the priority of the requirement. This is the normalized monthly average usage frequency, with a value range of 0 to 1. It is calculated by dividing the monthly average number of calls corresponding to the operation by the maximum monthly average number of calls for all operations. The efficiency coefficient of the operation path calculated in S2. To allow for efficiency improvements, the value range is 0 to 1; The percentage of users affected by demand, with a value ranging from 0 to 1, is calculated by dividing the number of users benefiting from the demand by the total number of active users in the system. The normalized technical implementation cost, with a value ranging from 0 to 1, is obtained through a comprehensive evaluation based on functional coupling, the proportion of reusable components, and code complexity. These are weighting coefficients, all ranging from 0 to 1, with the sum of all coefficients being 1. The default value can be adjusted based on the company's business strategy; the preferred default value is... ;

[0086] S33. Automatic generation of structured requirements specifications:

[0087] Based on the priority ranking results, a structured requirement specification document is automatically generated, which is divided into two main modules corresponding to the dual-path execution: the first type is the growth and expansion requirement list, which clarifies the requirement content, priority, target user group, and execution requirements; the second type is the optimization and contraction requirement list, which clarifies the optimization content, priority, scope of impact, and execution requirements. After the requirement specification document is generated, it is synchronized to the dual-path execution stage of S4.

[0088] S4, Adaptive optimization execution with dual parallel routes.

[0089] like Figure 3 As shown, it inherits the structured requirement specifications output by S3, and is based on the atomic foundation built by S1. It executes two parallel routes: requirement-driven functional growth and expansion, and behavior-driven optimization and contraction, to achieve orderly growth and healthy contraction of the system. All execution actions are completed in a hot deployment manner, without the need for service interruption and without interrupting the normal business operation of the enterprise.

[0090] S41. Execution of the demand-driven functional growth and expansion path:

[0091] To address the growth and expansion needs of S3 identification and solve the problem of insufficient initial functionality of the minimum atomic system, the system achieves reasonable and orderly growth of system functions through three types of actions: functional combination, capability addition, and best practice replication. All newly generated functions follow the atomic encapsulation specification of S1 and are incorporated into the atomic capability library for unified management.

[0092] S411, High-frequency atomic operation sequence merging and execution:

[0093] To address the need for combining high-frequency atomic operation sequences, the fixed atomic capability combinations frequently executed by users are transformed into standardized aggregation functions, which are then precipitated as reusable capability units for the system. The specific execution steps are as follows:

[0094] Step 1, Feasibility Assessment of the Compound:

[0095] The system automatically analyzes the input-output dependencies, transaction consistency requirements, and exception handling strategies of each atomic capability in the target sequence to determine whether they can be chained into a single aggregate function. The evaluation dimensions and passing criteria must simultaneously meet the following three requirements:

[0096] 1. Data flow feasibility: The output parameter set of the preceding atomic capability can completely cover the required input parameter set of the subsequent atomic capability, and the parameter type and data format are perfectly matched, with no missing parameters or format incompatibility issues;

[0097] 2. Transaction consistency: The transaction management logic of all atomic capabilities within the sequence can be uniformly converged, all operations can be included in the same global transaction, supporting unified commit and unified rollback, and there are no conflicts in transaction isolation levels;

[0098] 3. Anomaly Handling Compatibility: The anomaly level and handling logic of all atomic capabilities within the sequence can be unified. When an anomaly is triggered, a complete link rollback can be executed without any conflict in anomaly handling logic.

[0099] Step 2: Automated aggregation service code generation:

[0100] Based on the standardized APIs of atomic capabilities involved in the target sequence, aggregation service code is automatically generated, supporting multi-language code generation. The generated code fully includes parameter mapping, input validation, exception handling, idempotency guarantees, and end-to-end operation logging capabilities. The specific implementation logic is as follows:

[0101] 1. Extract the standardized interface definitions of each atomic capability in the target operation sequence, and construct the interface call chain according to the execution order of the operation sequence;

[0102] 2. Based on preset multilingual code templates, the main logic code for link connection is automatically generated. The code templates include standardized input parameter reception, parameter pass-through, and interface call order control logic.

[0103] 3. Automatically supplements the corresponding standardized code segments, including input validity verification code, exception capture and end-to-end rollback code, idempotency control code based on unique request identifier, and end-to-end operation log embedding code;

[0104] 4. Generate service code packages that can be directly compiled and deployed;

[0105] Step 3, Dynamic Hot Deployment of Functions:

[0106] The newly generated compound functions are dynamically registered to the system kernel through the service mesh, automatically generating function description documents, permission configuration items and interface entry points. They are then synchronized to the front end through the dynamic interface assembly mechanism described in S12, making them immediately available to users without the need for service interruption and version release.

[0107] Step 4: Precipitation of atomic abilities for chemical combination:

[0108] A preset observation period is set for the combination function. The observation period can be determined according to the enterprise's business iteration rhythm, with a baseline default of 2 calendar months. For agile iteration mode (iteration cycle ≤ 2 weeks), it is shortened to 1 month, and for waterfall development mode (iteration cycle ≥ 1 month), it is extended to 3 months. If the average monthly usage rate reaches the preset usage rate threshold within the observation period and user satisfaction meets the standard, the system automatically solidifies it into a new atomic capability and includes it in the atomic capability library. It can be used for subsequent combination and recombination. The usage rate threshold can be determined according to the importance of the business. The baseline default is an average monthly usage rate ≥ 30%, which is lowered to 20% for core business combination functions and raised to 40% for peripheral auxiliary functions. The average monthly usage rate = number of users calling in the current month / total number of target users. The user satisfaction rating adopts a 5-point scale and must simultaneously meet the following conditions: average score ≥ 4.0 points, effective rating sample ≥ 10% of the target user group, and negative rating (≤ 2 points) percentage < 5%.

[0109] Step 5: End-to-end traceability management:

[0110] The mapping relationship between chemical combination function and original atomic ability is permanently preserved, and chemical combination function can be disassembled, adjusted, or taken offline at any time. After the chemical combination function is disassembled, the original atomic ability is not affected in any way and can continue to operate independently or participate in other combination chemical combinations.

[0111] S412, Execution of adding atomic capabilities to fill business gaps:

[0112] To address business needs that cannot be covered by existing atomic capabilities, a non-intrusive extension of system capabilities is achieved. The specific steps are as follows:

[0113] The first step is to standardize the gap requirements. The system automatically transforms the identified capability gaps into standardized atomic capability requirement specifications, which clarify the business scenarios, input and output specifications, performance requirements, and target user groups.

[0114] The second step is development and integration. Standardized requirements are synchronized to the development team. The newly developed atomic capabilities follow the S1 encapsulation specification and are hot-deployed into the system through a dynamic registration mechanism without modifying the system kernel.

[0115] The third step is gray-scale verification and full release. New atomic capabilities are first gray-scale opened to the target user group triggered by the demand. The proportion of gray-scale users can be determined according to the size of the target user group. The usage rate, operation efficiency and user satisfaction are continuously monitored. After all indicators reach the preset full release standard, the full release is included in the system atomic capability library.

[0116] The fourth step is performance tracking. A preset observation period is set for newly connected atomic capabilities. If the usage rate is lower than the threshold within the observation period, a gradual degradation process to optimize the shrinkage route is automatically triggered to avoid the accumulation of invalid functions.

[0117] S413, Implementation of Cross-Group Best Practice Replication and Promotion:

[0118] The efficient operation sequences of individual users are combined into general functions and promoted throughout the enterprise, realizing the automatic accumulation and dissemination of best practices. The specific execution steps are as follows:

[0119] The first step is efficient sequence recognition. The system automatically identifies the operation sequence of a certain department or role. Its efficiency coefficient is significantly higher than the average level of the entire enterprise, and the business scenarios are applicable to the entire enterprise.

[0120] The second step is generalization and adaptation, which involves generalizing the efficient serialization into a general function and adapting it to the permissions and business rules of different departments and roles.

[0121] The third step is to promote the feature in batches, push it to the target user groups, continuously monitor usage and optimization results, and complete the best practice for enterprise-wide promotion.

[0122] S42. Behavior-driven optimized shrinkage route execution:

[0123] To address the optimization and shrinkage requirements of S3 identification and solve the problems of system function expansion and process inefficiency, the system achieves autonomous slimming through three types of actions: gradual function decommissioning, redundant function merging, and inefficient process reconstruction. This achieves a dynamic balance with the growth path. All actions are implemented based on the atomic and independent control capabilities of S1 and do not affect the normal operation of other functions.

[0124] S421. Implementation of progressive feature decommissioning based on the forgetting curve:

[0125] For low-frequency redundant atomic capabilities or chemical functions, automated decommissioning management is implemented throughout their entire lifecycle. The specific steps are as follows:

[0126] The first step is to model the forgetting curve, maintaining a unique forgetting curve for each atomic ability and chemical function, and quantifying the degree of forgetting. The calculation formula is as follows:

[0127] ;

[0128] In the formula, The value range represents the degree of functional forgetting. The higher the value, the higher the level of functional forgetting, and the deeper the corresponding stage of degradation. The current system time; This is the time when the function was last used; The base time period can be adjusted according to the characteristics of the enterprise's business; This is the function type coefficient, with a value range of 0 to 1. Core business functions have a value of 0.8 to 0.9, peripheral business functions have a value of 0.2 to 0.5, and non-core auxiliary functions have a value of 0 to 0.2. The more core the function, the higher the coefficient, and the slower the forgetting rate. The department's forgetting rate coefficient ranges from [0.5, 2]. Higher staff turnover and poorer business stability result in a higher coefficient and a faster forgetting rate. The calculation formula is as follows:

[0129] ;

[0130] In the formula, 1.0 is the general benchmark coefficient for enterprises, and 0.6 and 0.4 are the impact weights of employee turnover rate and business stability, respectively (the sum of the weights is 1, and enterprises can adjust them as needed). As a factor affecting staff turnover, Both are factors affecting business stability, and their values ​​range from [0.5, 2.5]. When the calculated result exceeds the range of [0.5, 2.5], the extreme values ​​are taken according to the upper and lower limits of the range.

[0131] Quantitative rules: Classified by departmental staff turnover rate over the past 90 days ( Values ​​are as follows: ≤5% = 0.5, 5%~10% = 0.8, 10%~20% = 1.2, 20%~30% = 1.8, ≥30% = 2.5;

[0132] Quantitative Rules: The average score of the frequency of core process changes and the fluctuation range of core business indicators in the past 90 days (out of 100) is used for tiered scoring: ≥90 points = 0.5, 75~90 points = 0.8, 60~75 points = 1.2, 40~60 points = 1.8, <40 points = 2.5; among them, no process changes = 100 points, and each change deducts 20 points; indicator fluctuation ≤5% = 100 points, and each fluctuation exceeding 5% deducts 15 points, with both scores having a minimum of 0 points.

[0133] The second step is progressive degradation execution, setting configurable time thresholds T1 to T4, which can be adjusted according to the enterprise's business characteristics. Simultaneously, a corresponding forgetting threshold is set for each degradation stage. When the function simultaneously meets the trigger conditions of usage interval and forgetting level, the corresponding degradation action is executed. The two conditions are logically ANDed. The specific stage rules are as follows:

[0134] 1. In stage 0, the function is active, the interval between function usage is less than the threshold T1, and the degree of forgetting is less than the corresponding threshold. The function is displayed normally and fully available.

[0135] 2. Stage 1 is a dormancy warning. The function usage interval is greater than or equal to T1, and the forgetting level is greater than or equal to the corresponding threshold. The function entry is moved to the secondary menu, usage prompts are added, and the core entry is retained. In this embodiment, T1 = 45 days, and D ≥ 0.3.

[0136] 3. Stage 2 is a cold storage state, where the function usage interval is greater than or equal to T2, and the degree of forgetting is greater than or equal to the corresponding threshold. The function is only visible during global search and no longer occupies foreground navigation resources, stopping unnecessary resource occupation. In this embodiment, T2 = 90 days and D ≥ 0.6.

[0137] 4. In stage 3, the function is in the archived state. The interval between function usages is greater than or equal to T3, and the degree of forgetting is greater than or equal to the corresponding threshold. The function is officially taken offline, and the corresponding data is archived and backed up. Users can apply for temporary restoration of use. In this embodiment, T3 = 180 days, and D ≥ 1.2.

[0138] 5. Stage 4 is the retirement state, where the interval between function usage is greater than or equal to T4, and the degree of forgetting is greater than or equal to the corresponding threshold. The function is completely removed, the corresponding code and container resources are all recycled, and the data is archived and deleted according to the enterprise rules. In this preferred embodiment, T4 = 270 days, and D ≥ 2.0.

[0139] The third step is to implement tiered offline management. For the combined functions generated by merging, two offline modes are provided: overall offline, which means directly offlineing the entire combined function, retaining the original atomic capabilities at the bottom layer, and not affecting other functions that reuse the same atom; and disassembly offline, which means disassembling the combined function into the original atomic capabilities, offlineing the atoms with a usage rate of 0, retaining the frequently reused atoms, and adding them to the atomic capability library for continued use.

[0140] The fourth step is the recall and recovery mechanism. When a user needs a feature that has been taken offline or archived, the recovery process is automatically triggered: First, the user's needs are verified by asking them about their usage scenarios and requirements through a pop-up window to avoid accidental recovery; then, a gradual recovery is performed, first restoring the core capabilities of the feature and gradually activating peripheral features as needed; finally, usage pattern learning is performed, continuously tracking the usage patterns of the feature after recovery. If the feature enters a forgetting cycle again, the time threshold for triggering the next degradation can be adjusted accordingly to prevent ineffective features from occupying resources for a long time.

[0141] S422. Merging and integrating duplicate / overlapping functions:

[0142] For atomic capabilities or combined functions with highly overlapping business scenarios and redundant functions, non-intrusive merging and integration can be achieved. The specific steps are as follows:

[0143] The first step is to identify redundant functions. Through the association rule algorithm and user behavior analysis described in S31, multiple functions with highly overlapping business scenarios and operation logic are identified, and the usage rate, user satisfaction, and resource consumption data of each function are statistically analyzed.

[0144] The second step is to automatically generate a merging scheme. Based on the atomic composition of the functions, the optimal merging scheme is automatically generated, retaining frequently used capabilities and eliminating redundant and duplicate capabilities. The merged functions follow the S1 atomic capability encapsulation specification.

[0145] The third step is hot update merging. The merged functions are dynamically launched through a hot-swappable architecture. The entry point of the original function is automatically redirected to the new function. A preset transition period is set, during which the temporary entry point of the original function is retained.

[0146] The fourth step is to remove redundant functions. After the transition period, verify that the usage rate and satisfaction of the merged functions meet the standards, gradually remove the original redundant functions, and reclaim system resources.

[0147] S423, Compression and Restructuring of Inefficient Processes and Paths:

[0148] To address user bypass operations and the sequential bottlenecks in the approval process, the process is optimized through compression, merging, and parallelization. The specific steps are as follows:

[0149] The first step is to mine inefficient paths by extracting high-frequency detour operation sequences and serial approval bottlenecks from operation logs and approval flow logs, identifying necessary nodes and optimizable nodes, and calculating the average processing time and data dependencies of each node.

[0150] The second step is to optimize the feasibility analysis. For the approval process, we will check whether there is data dependency between serial nodes. If there is no data dependency, we will automatically switch to parallel approval. If there is a dependency, we will assess whether partial parallelism can be achieved through a pre-approval mechanism. For the operation detour path, we will assess whether the operation path can be compressed through functional combination and interface optimization.

[0151] The third step is hot process updates. Through the dynamic deployment capabilities of the BPM engine (Business Process Management Engine), the original process is replaced with the optimized process definition. The running process is not affected and continues to execute according to the original process definition. The new process is only applied to newly initiated applications. The optimization of the operation path is updated synchronously through the interface dynamic assembly mechanism without the need for service interruption.

[0152] The fourth step is the rollback mechanism, which continuously monitors the performance of the optimized process. If any abnormalities occur, the process is automatically rolled back to the original process to avoid affecting the normal operation of the business.

[0153] In some embodiments, hot deployment includes version verification, dependency constraint verification, canary release and rollback mechanisms: assigning version numbers and dependency lists to newly added / changed atomic capabilities or compound functions; enabling and monitoring error rates, latency and resource consumption in the canary user group before release; automatically rolling back to the previous stable version when the monitored indicators exceed the threshold; and using equal control and compensation transactions to ensure data consistency in failure scenarios for compound functions involving cross-service calls.

[0154] Through the detailed description of the above embodiments, the method for constructing an Internet-based enterprise service application system of the present invention, by building a minimal atomic system foundation, provides a flexible and dynamically expandable underlying architecture support for the enterprise service system, breaking the rigid limitations of traditional stable architectures; through the full collection and semantic parsing of user operation carbon footprints, it achieves the accurate transformation from discrete operational behaviors to structured business requirements, providing objective and quantitative decision-making basis for the system's entire lifecycle iteration; through the design of a unified demand decision center and a dual-path adaptive optimization mechanism, it achieves a dynamic balance between the orderly growth and healthy contraction of system functions, ensuring the continuous adaptation of the system architecture to the enterprise's business development and digital transformation pace. The present invention can complete functional iteration without service downtime and achieve system adaptive optimization without extensive manual intervention, providing a complete and implementable technical solution for the construction and iteration of service systems throughout the entire lifecycle of enterprise digital transformation.

[0155] The above formulas are all dimensionless calculations, and the preset parameters in the formulas should be set by those skilled in the art according to the actual situation.

[0156] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0157] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0158] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0159] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0160] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0161] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0162] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for constructing an Internet-based enterprise service application system, characterized in that, The method flow is as follows: S1. Construct the foundation of the minimum atomic system, conduct architectural analysis and business capability decomposition of the target enterprise's full business domain, complete the standardized encapsulation of the minimum granularity atomic capabilities, form an atomic capability library that can be independently orchestrated, build a fully dynamic microkernel architecture that retains only basic management and control capabilities and supports hot-swappable access to business functions, and deploy the initial minimum feasible baseline system based on the atomic capabilities necessary for the enterprise's core business. S2. Collect and semantically analyze user operation carbon footprints. Collect user operation behavior data and corresponding context information in all dimensions without being aware of the user. Transform the discrete raw operation logs into structured business intent data. At the same time, generate standardized tags for users based on the analysis results and complete user group segmentation. Output structured data that can be used for demand identification. S3. Establish a unified demand decision center, synchronously identify two types of demands—functional growth and expansion, and optimization and contraction—based on semantic parsing results, construct a multi-dimensional priority scoring model to uniformly quantify and sort all demands, and automatically generate corresponding structured demand specifications. S4. Based on the aforementioned atomized system base, dual-path adaptive optimization is performed in parallel, involving demand-driven functional growth and expansion as well as behavior-driven optimization and contraction. All optimization actions are completed using a hot deployment method.

2. The method for constructing an Internet-based enterprise service application system according to claim 1, characterized in that, In step S1, the atomic capability is decomposed to the smallest indivisible business action node. After verifying the business boundary of the node, the corresponding execution logic and data interaction rules are matched to complete the initial encapsulation of the atomic capability. The encapsulated atomic capability has independent containerized deployment, independent start and stop management, standardized input and output interfaces, clear resource consumption boundaries, and full-link traceability.

3. The method for constructing an Internet-based enterprise service application system according to claim 2, characterized in that, The fully dynamic microkernel architecture predefines the standardized mapping relationship between interface element types and atomic capability interfaces. It can retrieve the corresponding list of accessible functions according to user permission configuration, group by business domain, and complete the layout and display of interface elements according to the function usage priority. The addition, deletion and modification of functions trigger the interface to automatically adapt synchronously.

4. The method for constructing an Internet-based enterprise service application system according to claim 1, characterized in that, In step S2, the collected operational behavior data includes explicit business operation data, implicit behavior preference data, cross-system off-site operation data, and operation context information. After all collected data is standardized in format, it is stored in the system log library as a data source for semantic parsing.

5. The method for constructing an Internet-based enterprise service application system according to claim 4, characterized in that, In step S2, semantic parsing first needs to construct a standardized mapping relationship between operation sequences and business intentions, complete the matching and transformation of discrete operation sequences into business intentions, then perform cluster statistics on the matched business intentions, and at the same time calculate the efficiency coefficient of the corresponding operation path and mark inefficient detour operations.

6. The method for constructing an Internet-based enterprise service application system according to claim 1, characterized in that, In step S3, the identified functional growth and expansion requirements include high-frequency atomic operation sequence combination requirements, business capability coverage gap requirements, and best practice replication requirements; the identified optimization and contraction requirements include low-frequency redundant function reduction requirements, duplicate and overlapping function merging requirements, and inefficient process optimization requirements.

7. The method for constructing an Internet-based enterprise service application system according to claim 6, characterized in that, The identification of duplicate and overlapping functions is achieved by taking the function call sequence within a single user session as a transaction set, iteratively mining frequent itemsets in the transaction set and generating association rules between functions, and calculating the function redundancy based on the support and confidence of the bidirectional association rules to complete the determination of duplicate and overlapping functions.

8. The method for constructing an Internet-based enterprise service application system according to claim 1, characterized in that, The execution of the functional growth and expansion path, for the high-frequency atomic operation sequence combination requirements, first evaluates the data flow feasibility, transaction consistency and exception handling compatibility of atomic capabilities within the sequence. After the evaluation is passed, the aggregate service code is automatically generated to complete the dynamic hot deployment and full-link traceability management of the function.

9. The method for constructing an Internet-based enterprise service application system according to claim 1, characterized in that, The optimized shrinkage route execution, for low-frequency redundant functions, maintains a dedicated forgetting curve for each function and quantifies the degree of forgetting, sets multi-stage progressive degradation actions according to the function usage interval and the degree of forgetting, and configures corresponding function recovery mechanisms and hierarchical offline management rules.

Images