Enterprise project management method and device, electronic equipment and storage medium

Abstract

The invention provides an enterprise project management method and device, electronic equipment and a storage medium, and belongs to the technical field of enterprise informatization management, in the method, cross-module real-time data synchronization is achieved through a unified data bus (UDB), data islands are eliminated, interaction with an external system is achieved through a protocol adapter, and the system reliability is improved. The interface adaptation efficiency is greatly improved, in addition, related business services call the target model to make decisions, that is, the target scheme corresponding to the business event is determined, quantitative support is achieved, in addition, the micro-service architecture is adopted, and expansibility and real-time performance are good.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise information management, and in particular to an enterprise project management method, device, electronic equipment and storage medium. Background Art

[0002] Current enterprise project management systems face multiple technical bottlenecks in complex business scenarios, specifically the following core issues: Data silos and failure of dynamic collaboration: Traditional systems use a modular, independent design, with data from each stage (pre-sales, project establishment, procurement, etc.) stored in separate databases and lacking unified semantic mapping rules. Inefficient multi-system interface adaptation: Enterprise external systems often utilize heterogeneous protocols. Specifically, these heterogeneous protocols manifest themselves in significant differences in the communication protocols used by different systems. For example, Web Services use XML for data exchange, while RESTful APIs tend to use JSON. Some systems may also use other formats such as binary transmission. These different protocols increase the complexity of interactions between systems and can also present challenges within the same business entities. Specifically, when connecting data from one system to another, it may be found that the fields corresponding to the same business entities in the two systems match less than 40% of the time. This means that most fields require some form of conversion or mapping. Traditional solutions require writing separate adaptation logic for each paired system, which not only consumes significant development resources but also requires manual code adjustments whenever either party upgrades its protocol version, further increasing maintenance costs.

[0003] Decision-making models lack quantitative support: Existing systems still rely on manual experience in key links such as risk assessment, supplier selection, and resource optimization.

[0004] Inadequate scalability and real-time performance: When traditional software architectures (such as monolithic designs) need to further expand the system to adapt to changes in enterprise size, costs rise rapidly, and this increase is not linear. This is because traditional monolithic architectures do not support good horizontal scalability (i.e., distributing the load by adding more servers), but instead rely on expensive hardware upgrades. Furthermore, to alleviate the pressure in high-concurrency scenarios, asynchronous processing and simplified computational logic may be employed. However, this only provides temporary relief and does not fundamentally resolve performance issues, resulting in increased response latency in high-concurrency scenarios.

[0005] In summary, traditional enterprise project management has technical problems such as data silos, inefficient multi-system interface adaptation, lack of quantitative support for decision-making models, and insufficient scalability and real-time performance. Summary of the Invention

[0006] In view of this, the purpose of the present invention is to provide an enterprise project management method, device, electronic device and storage medium to alleviate the technical problems of traditional enterprise project management, such as data silos, inefficient multi-system interface adaptation, lack of quantitative support for decision models, and insufficient scalability and real-time performance.

[0007] In a first aspect, an embodiment of the present invention provides an enterprise project management method, which is applied to an enterprise project management system. The method includes: The first module of the enterprise project management system publishes a business event, and distributes the business event to the second module via a unified data bus; The target service in the second module calls the target model to determine the target solution corresponding to the business event; The second module uses a protocol adapter to perform protocol conversion on the target solution, and sends the target solution after protocol conversion to an external system, so as to obtain, through the protocol adapter, a contract text solution returned by the external system for the target solution after protocol conversion; The target service in the second module invokes an NLP model to perform a dual-track risk analysis on the contract text scheme to obtain risk information of the contract text scheme, wherein the dual-track risk analysis includes: a risk analysis method based on a preset rule base and a risk analysis method using a deep learning model; The second module triggers a rolling time domain optimization algorithm based on the risk information to dynamically adjust the target plan to obtain an adjusted target plan.

[0008] Furthermore, distributing the business event to the second module via the unified data bus includes: Using a semantic standardization engine to perform standardization processing on the business event to obtain a standardized business event; The standardized business events are distributed to the second module via the unified data bus.

[0009] Furthermore, the second module uses a protocol adapter to perform protocol conversion on the target solution, including: Performing protocol conversion on the target solution using a protocol conversion matrix generated by an adversarial network to obtain a target solution after a first protocol conversion; Performing mapping conversion on the first field in the target solution after the first protocol conversion according to the protocol template library to obtain a target solution after the second protocol conversion, wherein the first field is a field that matches the field mapping rules for each pair of systems stored in the protocol template library; Performing fuzzy matching on the second field in the target solution after the second protocol conversion, and performing mapping conversion on the matched second field to obtain a target solution after the third protocol conversion, wherein the second field in the target solution after the second protocol conversion is a field other than the first field; Performing type identification on the unmatched second field in the target solution after the third protocol conversion, and performing inference conversion based on the identified type to obtain a target solution after the fourth protocol conversion; Manually review the unconverted fields in the target solution after the fourth protocol conversion to obtain the target solution after the protocol conversion.

[0010] Furthermore, the method further comprises: Performing a protocol conversion quality evaluation on the target solution after the protocol conversion to obtain the protocol conversion quality, wherein the protocol conversion quality evaluation includes: syntax layer protocol conversion quality evaluation, semantic layer protocol conversion quality evaluation and performance indicator protocol conversion quality; The protocol template library and / or the adversarial network are updated according to the protocol conversion quality, and / or the target resources are elastically expanded or contracted.

[0011] Furthermore, the target service in the second module calls the NLP model to perform a dual-track risk analysis on the contract text solution, including: Performing text preprocessing on the contract text scheme to obtain contract terms; Detecting specific risks corresponding to the contract terms based on a pre-set rule base; Use deep learning models to predict the risks of the contract terms and obtain potential risks; The explicit risks and the potential risks are combined to obtain risk information of the contract text solution.

[0012] Furthermore, the second module triggers a rolling horizon optimization algorithm based on the risk information to dynamically adjust the target solution, including: When the risk information is a high-risk clause or meets the dynamic adjustment conditions, the target solution is dynamically adjusted by adopting key parameter clipping, constraint relaxation, and parallel distributed solution to obtain the adjusted target solution.

[0013] Furthermore, the dynamic adjustment conditions include: the price fluctuation of key resources is greater than a threshold, the supplier is blacklisted, and the task progress deviation is greater than a preset threshold.

[0014] In a second aspect, an embodiment of the present invention further provides an enterprise project management device, which is applied to an enterprise project management system. The device includes: A publishing and distribution unit, configured to publish a business event to the first module of the enterprise project management system and distribute the business event to the second module via a unified data bus; A determination unit, configured for the target service in the second module to call a target model to determine a target solution corresponding to the business event; a protocol conversion unit configured to cause the second module to perform protocol conversion on the target solution using a protocol adapter, and to send the target solution after protocol conversion to an external system, so as to obtain, through the protocol adapter, a contract text solution returned by the external system for the target solution after protocol conversion; a dual-track risk analysis unit, configured to cause the target service in the second module to invoke an NLP model to perform a dual-track risk analysis on the contract text scheme to obtain risk information of the contract text scheme, wherein the dual-track risk analysis includes: a risk analysis method based on a preset rule base and a risk analysis method using a deep learning model; A dynamic adjustment unit is used for the second module to trigger a rolling time domain optimization algorithm based on the risk information to dynamically adjust the target plan to obtain an adjusted target plan.

[0015] In a third aspect, an embodiment of the present invention further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any one of the methods described in the first aspect when executing the computer program.

[0016] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to execute any method described in the first aspect above.

[0017] In an embodiment of the present invention, an enterprise project management method is provided, which is applied to an enterprise project management system. The method includes: a first module of the enterprise project management system publishes a business event and distributes the business event to a second module through a unified data bus; a target service in the second module calls a target model to determine a target solution corresponding to the business event; the second module uses a protocol adapter to perform protocol conversion on the target solution, and sends the target solution after protocol conversion to an external system to obtain a contract text solution returned by the external system for the target solution after protocol conversion through the protocol adapter; the target service in the second module calls an NLP model to perform a dual-track risk analysis on the contract text solution to obtain risk information of the contract text solution; the second module triggers a rolling time domain optimization algorithm based on the risk information to dynamically adjust the target solution to obtain an adjusted target solution. From the above description, it can be seen that in the enterprise project management method of the present invention, real-time data synchronization across modules is achieved through the unified data bus (UDB), data silos are eliminated, and interaction with external systems is achieved through the protocol adapter, greatly improving the interface adaptation efficiency. In addition, related business services call the target model to make decisions, that is, determine the target plan corresponding to the business event, and realize quantitative support. In addition, a microservice architecture is also adopted, which has good scalability and real-time performance, alleviating the technical problems of traditional enterprise project management such as data silos, low efficiency of multi-system interface adaptation, lack of quantitative support for decision-making models, and insufficient scalability and real-time performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 A flowchart of an enterprise project management method provided by an embodiment of the present invention; Figure 2 A schematic diagram of risks after merging provided by an embodiment of the present invention; Figure 3 A schematic diagram of an enterprise project management device provided by an embodiment of the present invention; Figure 4 A schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] Traditional enterprise project management suffers from data silos, inefficient multi-system interface adaptation, lack of quantitative support for decision-making models, and insufficient scalability and real-time performance.

[0022] Based on this, in the enterprise project management method of the present invention, real-time data synchronization across modules is achieved through the unified data bus (UDB), eliminating data silos, and interaction with external systems is achieved through the protocol adapter, greatly improving the interface adaptation efficiency. In addition, related business services call the target model to make decisions, that is, determine the target plan corresponding to the business event, and realize quantitative support. In addition, a microservice architecture is also adopted, which has good scalability and real-time performance.

[0023] To facilitate understanding of this embodiment, an enterprise project management method disclosed in an embodiment of the present invention is first introduced in detail.

[0024] Example 1: According to an embodiment of the present invention, an embodiment of an enterprise project management method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0025] Figure 1 is a flow chart of an enterprise project management method according to an embodiment of the present invention. Figure 1 As shown, the method includes the following steps: Step S102: The first module of the enterprise project management system publishes a business event and distributes the business event to the second module via a unified data bus; In an embodiment of the present invention, the enterprise project management system includes a pre-sales module, a project establishment module, a procurement module, a subcontracting module, an inventory module, a financial module, and a multi-system interface collaboration module, enabling intelligent collaborative management of enterprise projects throughout their entire lifecycle. The first module may be a pre-sales module, and the second module may be a procurement module. However, the embodiments of the present invention do not impose specific limitations on the first and second modules.

[0026] It's important to note that all modules output data in a standardized JSON format, with mandatory business type tags. They use RabbitMQ / RocketMQ for inter-module communication, eliminating complex stream processing. Each module broadcasts in real time, achieving synchronized notifications across multiple modules within milliseconds.

[0027] Specifically, after a first module publishes a business event, it distributes the business event to a second module via a unified data bus. Specifically, the second module is the module that has subscribed to the business event. For example, if the pre-sales module publishes a purchase request event (i.e., a user submits a purchase request in the pre-sales module), the purchase request event is distributed via the unified data bus (UDB) to the subscribed procurement, inventory, and finance modules. Compared to traditional point-to-point calls (strong coupling (direct interface connection)), this UDB event-driven architecture offers loose coupling (each module only needs to connect to the UDB and subscribe to events). If new modules are added without modifying existing interfaces, the UDB event-driven architecture can dynamically register consumers (i.e., elastic scalability, allowing consumers to scale independently). This advantage is that expansion is time-efficient. In other words, in the UDB event-driven architecture of the present invention, adding new modules does not require modifying the pre-sales module code, a single module failure will not cause a cascading crash, and each module can be independently upgraded and iterated. Furthermore, the event flow implemented via the UDB exhibits atomicity: events are either delivered in full or discarded; sequentiality: correct business operation timing is guaranteed; and persistence: audit logs are maintained for all events. In addition, in traditional solutions, the data format of each module is customized, new fields need to be reconstructed, and parsing efficiency requires type inference. The data format of the solution of the present invention is standardized in JSON template, metadata can be added dynamically, and parsing efficiency is accelerated by predefined Schema.

[0028] Step S104: The target service in the second module calls the target model to determine the target solution corresponding to the business event; Specifically, if the procurement module (i.e., the second module, used here as an example) receives a procurement demand event (i.e., a business event) issued by the pre-sales module (i.e., the first module), it calls the supplier service (i.e., the target service), and the supplier service then calls the supplier graph model (i.e., the target model), filters the Top-N candidate set, and then calls the cost optimization model (i.e., the target model) to calculate the optimal solution (i.e., the target solution) among the Top-N candidate sets.

[0029] The embodiment of the present invention does not impose any specific restrictions on the above-mentioned target model, and it can be adjusted according to specific business scenarios. For example, when dealing with uncertainty problems such as supplier selection, a fuzzy comprehensive evaluation algorithm can be used; when dealing with multi-stage resource allocation problems, a dynamic programming algorithm can be used; when optimizing logistics path planning, a graph theory algorithm can be used; or a combination algorithm can be used. For example, in the business scenario of supplier selection, a combination algorithm of fuzzy evaluation and PageRank can be used (parameters to be considered include: quality score / historical default rate / price, and the specific output can be a recommended list of top 3 suppliers); in the business scenario of procurement batch planning, a combination algorithm of dynamic programming and linear regression can be used (parameters to be considered include: inventory cost / transportation rate / demand forecast, and the specific output can be the optimal procurement quantity and time node); in the business scenario of project resource scheduling, a combination algorithm of critical path method (CPM) and genetic algorithm can be used (parameters to be considered include: task dependency / resource constraint, and the specific output can be a resource Gantt chart and critical path identifier).

[0030] For example, the input demand is that 8,000 tons of X70 pipeline steel needs to be purchased with the quality standard of API 5L. After calling the supplier map model (i.e., the target model), 23 qualified suppliers are screened and the supplier ID, PageRank, and quality score are obtained. Then, the cost optimization model is further called to obtain the preset number of suppliers with the lowest cost from the 23 qualified suppliers.

[0031] In addition, the above process also supports incremental learning. For example, in the optimization of steel raw material procurement, the initial state is that the PageRank value of supplier A is 0.35, and the PageRank value of supplier B is 0.28. During the learning process, the incremental data is that supplier A has delayed delivery for three consecutive times. The PageRank value of supplier A is updated to 0.7×0.35, that is, the PageRank value of supplier A is penalized by the penalty coefficient, and the PageRank value of supplier B is updated to 0.1+0.28, that is, the PageRank value of supplier B is rewarded by competitive rewards. The final output result is: updated supplier ranking: B>A, and the order allocation ratio of A is automatically reduced.

[0032] Through the innovative architecture of "dynamic algorithm combination plus real-time incremental learning", this invention achieves business adaptability and automatic matching of the optimal algorithm combination; it achieves decision-making intelligence and sustainable evolution of decision-making models; it achieves resource economy and maximizes the use of computing resources.

[0033] Step S106: The second module uses a protocol adapter to perform protocol conversion on the target solution, and sends the target solution after protocol conversion to the external system, so as to obtain, through the protocol adapter, a contract text solution returned by the external system for the target solution after protocol conversion; The above-mentioned external systems can be legal systems, financial systems, audit systems, bidding systems, etc.

[0034] In step S108, the target service in the second module invokes the NLP model to perform a dual-track risk analysis on the contract text solution to obtain risk information of the contract text solution. The dual-track risk analysis includes: risk analysis based on a preset rule base and risk analysis using a deep learning model. In step S110 , the second module triggers a rolling time domain optimization algorithm based on the risk information to dynamically adjust the target solution to obtain an adjusted target solution.

[0035] The process of step S110 is described below with an example: In the steel procurement case, the demand input is that 2,000 tons of cold-rolled steel plates need to be purchased, and the deadline is 2023-08-30.

[0036] Supplier matching: Input 10 qualified suppliers; output ranking, supplier and comprehensive score are: 1, Baosteel, 92.5; 2, Ansteel, 88.3; 3, Shagang, 85.7.

[0037] After compliance verification (risk analysis), it was found that the Baosteel contract contained an "unconditional pass-through of raw material price increases" clause (high risk). The Ansteel solution was automatically downgraded to Ansteel. The final output was: Ansteel, contract_template was STEEL-PURCHASE-2023, and risk_notes indicated a price fluctuation clause requiring manual review.

[0038] In an embodiment of the present invention, an enterprise project management method is provided, which is applied to an enterprise project management system. The method includes: a first module of the enterprise project management system publishes a business event and distributes the business event to a second module through a unified data bus; a target service in the second module calls a target model to determine a target solution corresponding to the business event; the second module uses a protocol adapter to perform protocol conversion on the target solution, and sends the target solution after protocol conversion to an external system to obtain a contract text solution returned by the external system for the target solution after protocol conversion through the protocol adapter; the target service in the second module calls an NLP model to perform a dual-track risk analysis on the contract text solution to obtain risk information of the contract text solution; the second module triggers a rolling time domain optimization algorithm based on the risk information to dynamically adjust the target solution to obtain an adjusted target solution. From the above description, it can be seen that in the enterprise project management method of the present invention, real-time data synchronization across modules is achieved through the unified data bus (UDB), data silos are eliminated, and interaction with external systems is achieved through the protocol adapter, greatly improving the interface adaptation efficiency. In addition, related business services call the target model to make decisions, that is, determine the target plan corresponding to the business event, and realize quantitative support. In addition, a microservice architecture is also adopted, which has good scalability and real-time performance, alleviating the technical problems of traditional enterprise project management such as data silos, low efficiency of multi-system interface adaptation, lack of quantitative support for decision-making models, and insufficient scalability and real-time performance.

[0039] The above content briefly introduces the enterprise project management method of the present invention. The specific contents involved are described in detail below.

[0040] In an optional embodiment of the present invention, distributing the business event to the second module via the unified data bus specifically includes the following steps: (1) Using a semantic standardization engine to standardize business events and obtain standardized business events; (2) Distribute the standardized business events to the second module through the unified data bus.

[0041] Specifically, after the pre-sales module submits a JSON-formatted demand, the event is serialized and stored in a Kafka partition, where it is processed in parallel by consumer groups (response within milliseconds). During implementation, a semantic standardization engine is first used to standardize business events. This standardized business event ensures that the data flowing in the bus has unified semantics. After conversion is completed on the production side, the consumer side can directly use the standardized data. This means that the present invention achieves real-time data synchronization across modules based on the unified data bus (UDB). The semantic standardization engine eliminates data silos, meaning that demand and resource data flow in real time via the UDB (unified data bus), and semantic standardization ensures consistency.

[0042] Semantic standardization engine: defines the global metadata model (GMM) and uniformly maps the data of each module into a fixed format.

[0043] Unified Data Bus (UDB): Builds a real-time data stream pipeline based on Apache Kafka, supporting multi-module data transmission classified by business topic (Topic).

[0044] In an optional embodiment of the present invention, the second module uses a protocol adapter to perform protocol conversion on the target solution, specifically comprising the following steps: (1) Use the protocol conversion matrix generated by the adversarial network to perform protocol conversion on the target solution to obtain the target solution after the first protocol conversion; Specifically, the protocol conversion matrix is ​​used to address differences in protocol structures. It is generated through an adversarial network (GAN). This matrix is ​​triggered only when a new protocol type is first introduced, and it then generates a protocol conversion matrix. It primarily addresses macro-protocols. In the GAN architecture, the generator structure can be: protocol A structure → encoder → latent space → decoder → protocol B structure.

[0045] When training the adversarial network, the input is: 100,000 sets of protocol pairs (WSDL / OpenAPI / binary, etc.), and the loss function is: L = αLrecon + βLadv + γLstruct. Among them, the structural loss Lstruct ensures that key business elements are not lost.

[0046] The protocol conversion matrix generated by GAN can be shown as follows: Protocol A field 1 → Protocol B field 1, Protocol A field 2 → Protocol B fields 2+3, Protocol A field 3 → Protocol B field 3. It supports parallel processing of protocol conversion tasks such as HTTP / API / WebService.

[0047] (2) performing mapping conversion on the first field in the target solution after the first protocol conversion according to the protocol template library to obtain the target solution after the second protocol conversion, wherein the first field is a field that matches the field mapping rules of each pair of systems stored in the protocol template library; Specifically, the present invention adopts a three-level conversion for the protocol. The first level is template mapping. The protocol template library is first initialized: high-frequency field mapping is extracted from historical data and then stored in the protocol template library. That is, the preset protocol templates in the protocol template library are mapping rules for each pair of system-defined fields. Then, according to the protocol template library, the first field in the target scheme after the first protocol conversion is mapped and converted to obtain the target scheme after the second protocol conversion.

[0048] (3) performing fuzzy matching on the second field in the target solution after the second protocol conversion, and performing mapping conversion on the matched second field to obtain a target solution after the third protocol conversion, wherein the second field in the target solution after the second protocol conversion is a field other than the first field; Specifically, the second level of the three-level conversion is intelligent mapping, which includes fuzzy matching. This involves using the Levenshtein distance algorithm to calculate field name similarity. If the field name similarity reaches a preset threshold of 80%, automatic mapping is performed. Fuzzy matching can handle common field differences (such as capitalization, separators, and abbreviations), covering 60%-70% of common field mismatches.

[0049] (4) Identify the type of the unmatched second field in the target solution after the third protocol conversion, and perform inference conversion based on the identified type to obtain the target solution after the fourth protocol conversion; Specifically, intelligent mapping also includes type-driven analysis. For example, by identifying the types of ￥ and $ as currency, it can be mapped to price fields. Similarly, by identifying the yyy-mm-dd format as a date, it can be standardized to ISO8601. This conversion uses regular expressions for rapid pattern recognition, supports inference of 12 basic data types, and can handle approximately 15%-20% of unconventional fields.

[0050] (5) Manually review the unconverted fields in the target solution after the fourth protocol conversion to obtain the target solution after the protocol conversion.

[0051] Specifically, the third level of the three-level conversion is manual review (safety net mechanism), and the triggering condition is that the first two levels of processing have failed.

[0052] In the present invention, adaptive protocol conversion is achieved through "configuration templates and intelligent backup" to improve development efficiency; the specific implementation is divided into three levels: the first level preset templates can solve 80% of routine conversions, the second level field fuzzy matching and type drive handle common differences, and the third level manual review covers abnormal situations. During implementation, the adversarial network is triggered only when a new protocol type is first accessed; all data conversions must go through three levels of conversion; protocols with existing protocol conversion matrices can bypass GAN processing. The goal of the three-level conversion process is to solve field-level mapping, and the technologies used are rule engine, fuzzy matching, and type drive, which must be executed every time data is converted, and finally output a field mapping table.

[0053] It should be noted that the manual review results will automatically generate a new agreement template. For example, when similar fields appear more than three times, they will be automatically added to the agreement template library.

[0054] In an optional embodiment of the present invention, the method further comprises the following steps: (1) Perform protocol conversion quality evaluation on the target solution after protocol conversion to obtain protocol conversion quality, where the protocol conversion quality evaluation includes: syntax layer protocol conversion quality evaluation, semantic layer protocol conversion quality evaluation and performance indicator protocol conversion quality; (2) Update the protocol template library and / or adversarial network based on the quality of protocol conversion, and / or elastically scale the target resources.

[0055] Specifically, the system self-evolves based on protocol conversion quality indicators. This is primarily constructed from three dimensions: First, the syntax layer, using field mapping completeness (number of missing fields / total number of fields) and data type matching rate (number of type errors / total converted fields) to quantify basic conversion quality. For example, missing required fields trigger an emergency alert. Second, the semantic layer verifies the logical rationality of converted data using business rules. For example, if the purchase amount suddenly drops to zero, which is clearly unreasonable, the number of rule violations is quantified by dividing the total number of verifications. Finally, performance indicators are used. Latency and throughput anomalies may indicate decreased protocol adapter efficiency.

[0056] Specifically, a multi-level threshold design is used: hard errors (such as missing fields) directly trigger real-time updates to the protocol template library. Soft deviations (such as numerical fluctuations) accumulate to a threshold, triggering batch training (i.e., adversarial network training). If performance degrades, resource expansion is initiated. During implementation, when an abnormal-level hard error occurs, such as missing key fields or severe type mismatches, the process can be immediately blocked, triggering real-time rule updates and alerting operations and maintenance. When an abnormal-level soft deviation occurs, such as numerical fluctuations exceeding ±15% but not exceeding business red lines, an abnormality log is recorded. Ten cumulative occurrences trigger model retraining, and a diagnostic report is generated. When abnormal-level performance degrades, such as latency exceeding 200ms for five minutes or a 20% drop in TPS, container instances can be automatically scaled up (supporting horizontal scaling, microservice-based splitting, and containerized deployment, with automatic resource allocation adjusted based on load), switching to an alternative protocol path, and triggering performance optimization tasks.

[0057] When implemented, the comprehensive evaluation formula for protocol conversion quality can be ,in, Indicates the quality of protocol conversion. Indicates the quality of syntax layer conversion, Indicates the quality of semantic layer conversion, Indicates the quality of performance indicators, 、 、 All∈[0,1], ,default value , , .

[0058] Evaluation formula for syntax layer conversion quality: ,in, Indicates the quality of syntax layer conversion, Indicates the value of the i-th field of the source protocol, Indicates the corresponding field value of the target protocol. Indicates the field tolerance threshold, Indicates field weight; Evaluation formula for semantic layer conversion quality: ,in, Indicates the quality of semantic layer conversion, represents the accuracy of the j-th business rule, represents the recall rate of the j-th business rule; Evaluation formula for performance index quality: ,in, Indicates the quality of performance indicators, Indicates actual delay / requested delay, Indicates actual throughput / required throughput, represents the delay sensitivity coefficient; When the value is less than 0.8, the protocol template library is updated. The template library update strategy is as follows: ,in, represents the weight of the kth rule in the protocol template library, represents the learning rate, Indicates the quality of protocol conversion. Indicates the quality threshold; When <0.7, the adversarial network is retrained and the GAN network update strategy is: ,in, represents the original loss function, represents the quality adjustment coefficient, represents the generator output, represents the ideal conversion sample; When the value is less than 0.6, the target resources are elastically scaled up or down. The resource scaling decision is: ,in, Represents the single node processing capacity coefficient, when When it is greater than 0, expansion is triggered. When <0, shrinkage is triggered.

[0059] In an optional embodiment of the present invention, the target service in the second module calls the NLP model to perform a dual-track risk analysis on the contract text solution, specifically including the following steps: (1) Preprocess the contract text to obtain the contract terms; (2) Detecting specific risks corresponding to contract terms based on a preset rule base; (3) Use deep learning models to predict the risks of contract terms and obtain potential risks; (4) Combine explicit risks and potential risks to obtain risk information for the contract text plan.

[0060] Specifically, during text preprocessing, the contract text is segmented according to the legal clause structure, identifying the contract parties and key entities and removing non-normative content. The dual-track analysis includes: a rule engine that detects specific risks associated with contract clauses based on a preset rule base (keyword regular matching and a preset risk rule base, such as payment exceeding 60 days, unequal contractual liability, intellectual property ownership, etc.); and a deep learning model that uses fine-tuned BERT to identify potential risks in contract clauses (for example, input: contract clause (such as "delivery delayed due to force majeure is not considered a breach of contract"), output: corresponding risk label and confidence level). The dual-track results are confidence-weightedly fused to produce a classification report with risk levels (i.e., risk information). The fine-tuned BERT was trained using 10,000 historical contract annotations (normal and high-risk clauses). Its network architecture is a 4-layer distilled BERT (model size <50MB). After deployment, the CPU inference speed is <0.5 seconds per contract.

[0061] For example, taking the purchase contract returned by the legal system as an example, we extract the following clauses (such as "payment method" and "liability for breach of contract") through text preprocessing: extract the clauses in sections (such as "payment method" and "liability for breach of contract"), remove irrelevant symbols, and retain the legal entity (Party A / Party B). Through dual-track analysis, we find that "Payment cycle: final payment 90 days after order takes effect" → triggers the "payment cycle too long" risk, and the price fluctuation clause is out of control. Risks are merged into Figure 2 The results shown.

[0062] In an optional embodiment of the present invention, the second module triggers a rolling horizon optimization algorithm based on risk information to dynamically adjust the target solution, specifically including the following steps: When the risk information is a high-risk clause or meets the dynamic adjustment conditions, the target plan is dynamically adjusted by cutting key parameters, relaxing constraints, and solving in parallel and distributed manner to obtain the adjusted target plan.

[0063] Specifically, dynamic adjustment based on Rolling Horizon Optimization (RHO) triggers resource reallocation and generates an adjusted target plan.

[0064] A dynamic adjustment solution based on Rolling Horizon Optimization (RHO): key parameter tailoring (adjusting resources involved in the change, such as calculating only steel-related tasks when steel prices increase), constraint relaxation (allowing temporary breakthroughs in original restrictions, such as adjusting the budget limit), and parallel distributed solution (independent calculations by resource type, and merging the results).

[0065] Contingency Plan Content: Set trigger conditions. Meeting just one set triggers subsequent emergency procedures (i.e., dynamic adjustments). Examples of trigger conditions (dynamic adjustments) include: key resource price fluctuations exceeding a threshold (e.g., a 15% steel price increase), a supplier being blacklisted, or a task progress deviation exceeding 10%. Automated approval rules can be implemented based on business needs, such as: Cost increase <5% and within deadlines → immediate effect. Manual processing can also be added, such as: Cost increase >5% → email notification to the project manager for confirmation.

[0066] In an optional embodiment of the present invention, the dynamic adjustment conditions include: the price fluctuation of key resources is greater than a threshold, the supplier is included in the blacklist, and the task progress deviation is greater than a preset threshold.

[0067] The solution of this invention replaces complex algorithm coupling with an "event-driven + rule engine" approach, enabling self-learning capabilities. This involves automated rule generation, coupled with a closed-loop manual feedback loop to achieve self-learning within the architecture. This approach meets the actual needs of enterprises while avoiding complex AI training at a low cost. The invention also provides a visual decision-making platform with embedded risk warning and scenario simulation capabilities, offering multi-dimensional views such as risk heat maps, resource Gantt charts, and supplier topology maps.

[0068] This invention overcomes the problems of data fragmentation, response lag and manual dependence in traditional systems through the three pillars of multi-model fusion, interface intelligence and data collaboration, provides enterprises with a full-life cycle, adaptively optimized project management center, and promotes improved operational efficiency.

[0069] Example 2: The embodiment of the present invention further provides an enterprise project management device, which is mainly used to execute the enterprise project management method provided in the first embodiment of the present invention. The enterprise project management device provided in the embodiment of the present invention is specifically introduced below.

[0070] Figure 3 Schematic diagram of an enterprise project management device according to an embodiment of the present invention. Figure 3 As shown, the device mainly includes: a publishing and distribution unit 10, a determination unit 20, a protocol conversion unit 30, a dual-track risk analysis unit 40 and a dynamic adjustment unit 50, wherein: A publishing and distribution unit is used for publishing business events in the first module of the enterprise project management system and distributing the business events to the second module via a unified data bus; A determination unit, configured for the target service in the second module to call the target model to determine a target solution corresponding to the business event; A protocol conversion unit is configured to perform protocol conversion on the target solution by using a protocol adapter in the second module, and send the target solution after protocol conversion to an external system, so as to obtain, through the protocol adapter, a contract text solution returned by the external system for the target solution after protocol conversion; A dual-track risk analysis unit is used by the target service in the second module to call the NLP model to perform a dual-track risk analysis on the contract text solution to obtain risk information of the contract text solution. The dual-track risk analysis includes: risk analysis based on a preset rule base and risk analysis using a deep learning model; The dynamic adjustment unit is used for the second module to trigger the rolling time domain optimization algorithm based on risk information to dynamically adjust the target plan to obtain the adjusted target plan.

[0071] In an embodiment of the present invention, an enterprise project management device is provided, which is applied to an enterprise project management system. The device includes: a first module of the enterprise project management system publishes a business event and distributes the business event to a second module through a unified data bus; the target service in the second module calls a target model to determine a target solution corresponding to the business event; the second module uses a protocol adapter to perform protocol conversion on the target solution, and sends the target solution after protocol conversion to an external system to obtain a contract text solution returned by the external system for the target solution after protocol conversion through the protocol adapter; the target service in the second module calls an NLP model to perform a dual-track risk analysis on the contract text solution to obtain risk information of the contract text solution; the second module triggers a rolling time domain optimization algorithm based on the risk information to dynamically adjust the target solution to obtain an adjusted target solution. From the above description, it can be seen that in the enterprise project management device of the present invention, real-time data synchronization across modules is achieved through the unified data bus (UDB), data silos are eliminated, and interaction with external systems is achieved through the protocol adapter, and the interface adaptation efficiency is greatly improved. In addition, the relevant business services call the target model to make decisions, that is, determine the target plan corresponding to the business event, and realize quantitative support. In addition, a microservice architecture is also adopted, which has good scalability and real-time performance, alleviating the technical problems of traditional enterprise project management such as data silos, low efficiency of multi-system interface adaptation, lack of quantitative support for decision-making models, and insufficient scalability and real-time performance.

[0072] Optionally, the publishing and distributing unit is further configured to: perform standardization processing on the business events using a semantic standardization engine to obtain standardized business events; and distribute the standardized business events to the second module via a unified data bus.

[0073] Optionally, the protocol conversion unit is also used to: perform protocol conversion on the target scheme using the protocol conversion matrix generated by the adversarial network to obtain the target scheme after the first protocol conversion; perform mapping conversion on the first field in the target scheme after the first protocol conversion according to the protocol template library to obtain the target scheme after the second protocol conversion, wherein the first field is a field that matches the field mapping rules of each pair of systems stored in the protocol template library; perform fuzzy matching on the second field in the target scheme after the second protocol conversion, and perform mapping conversion on the matched second field to obtain the target scheme after the third protocol conversion, wherein the second field in the target scheme after the second protocol conversion is a field other than the first field; perform type identification on the unmatched second field in the target scheme after the third protocol conversion, and perform inference conversion based on the identified type to obtain the target scheme after the fourth protocol conversion; perform manual review on the unconverted fields in the target scheme after the fourth protocol conversion to obtain the target scheme after the protocol conversion.

[0074] Optionally, the device is also used to: perform protocol conversion quality evaluation on the target solution after protocol conversion to obtain protocol conversion quality, wherein the protocol conversion quality evaluation includes: syntax layer protocol conversion quality evaluation, semantic layer protocol conversion quality evaluation and performance indicator protocol conversion quality; update the protocol template library and / or adversarial network according to the protocol conversion quality, and / or elastically expand or shrink the target resources.

[0075] Optionally, the dual-track risk analysis unit is also used to: perform text preprocessing on the contract text scheme to obtain contract terms; detect explicit risks corresponding to the contract terms based on a preset rule base; use a deep learning model to predict risks on the contract terms to obtain potential risks; merge explicit risks and potential risks to obtain risk information of the contract text scheme.

[0076] Optionally, the dynamic adjustment unit is also used to: when the risk information is a high-risk clause or meets the dynamic adjustment conditions, dynamically adjust the target plan by using key parameter clipping, constraint relaxation, and parallel distributed solution to obtain an adjusted target plan.

[0077] Optionally, the dynamic adjustment conditions include: the price fluctuation of key resources is greater than a threshold, the supplier is blacklisted, and the task progress deviation is greater than a preset threshold.

[0078] The device provided in the embodiment of the present invention has the same implementation principle and technical effects as those in the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the device embodiment, reference can be made to the corresponding content in the aforementioned method embodiment.

[0079] like Figure 4As shown, an electronic device 600 provided in an embodiment of the present application includes: a processor 601, a memory 602 and a bus, wherein the memory 602 stores machine-readable instructions executable by the processor 601. When the electronic device is running, the processor 601 communicates with the memory 602 through the bus, and the processor 601 executes the machine-readable instructions to perform the steps of the above-mentioned enterprise project management method.

[0080] Specifically, the memory 602 and processor 601 can be general-purpose memories and processors, which are not specifically limited here. When the processor 601 runs the computer program stored in the memory 602, the enterprise project management method can be executed.

[0081] The processor 601 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by an integrated logic circuit of hardware in the processor 601 or by instructions in the form of software. The above-mentioned processor 601 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The various methods, steps, and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present application can be directly embodied as being executed by a hardware decoding processor, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium well-known in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in memory 602, and processor 601 reads the information in memory 602 and performs the steps of the above method in conjunction with its hardware.

[0082] Corresponding to the above-mentioned enterprise project management method, an embodiment of the present application also provides a computer-readable storage medium, which stores machine-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to execute the steps of the above-mentioned enterprise project management method.

[0083] The enterprise project management device provided in the embodiment of the present application can be specific hardware on the device or software or firmware installed on the device. The device provided in the embodiment of the present application, its implementation principle and the technical effect produced are the same as those in the aforementioned method embodiment. For the sake of brief description, where the device embodiment is not mentioned, reference can be made to the corresponding content in the aforementioned method embodiment. Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can all refer to the corresponding processes in the aforementioned method embodiment, and will not be repeated here.

[0084] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interface, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0085] For another example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of the boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0086] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0087] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0088] If the functions are implemented in the form of 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 the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling an electronic device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the enterprise project management method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0089] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0090] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The scope of protection of the present application is not limited thereto. Although the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-mentioned embodiments within the technical scope disclosed in the present application, or perform equivalent replacements for some of the technical features thereof. However, these modifications, changes, or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application. They should all be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. An enterprise project management method, characterized in that: Applied to an enterprise project management system, the method includes: The first module of the enterprise project management system publishes a business event, and distributes the business event to the second module via a unified data bus; The target service in the second module calls the target model to determine the target solution corresponding to the business event; The second module uses a protocol adapter to perform protocol conversion on the target solution, and sends the target solution after protocol conversion to an external system, so as to obtain, through the protocol adapter, a contract text solution returned by the external system for the target solution after protocol conversion; The target service in the second module invokes an NLP model to perform a dual-track risk analysis on the contract text scheme to obtain risk information of the contract text scheme, wherein the dual-track risk analysis includes: a risk analysis method based on a preset rule base and a risk analysis method using a deep learning model; The second module triggers a rolling time domain optimization algorithm based on the risk information to dynamically adjust the target plan to obtain an adjusted target plan.

2. The method according to claim 1, characterized in that Distributing the business event to the second module through the unified data bus includes: Using a semantic standardization engine to perform standardization processing on the business event to obtain a standardized business event; The standardized business events are distributed to the second module via the unified data bus.

3. The method according to claim 1, characterized in that The second module uses a protocol adapter to perform protocol conversion on the target solution, including: Performing protocol conversion on the target solution using a protocol conversion matrix generated by an adversarial network to obtain a target solution after a first protocol conversion; Performing mapping conversion on the first field in the target solution after the first protocol conversion according to the protocol template library to obtain a target solution after the second protocol conversion, wherein the first field is a field that matches the field mapping rules for each pair of systems stored in the protocol template library; Performing fuzzy matching on the second field in the target solution after the second protocol conversion, and performing mapping conversion on the matched second field to obtain a target solution after the third protocol conversion, wherein the second field in the target solution after the second protocol conversion is a field other than the first field; Performing type identification on the unmatched second field in the target solution after the third protocol conversion, and performing inference conversion based on the identified type to obtain a target solution after the fourth protocol conversion; Manually review the unconverted fields in the target solution after the fourth protocol conversion to obtain the target solution after the protocol conversion.

4. The method according to claim 3, characterized in that The method further comprises: Performing a protocol conversion quality evaluation on the target solution after the protocol conversion to obtain the protocol conversion quality, wherein the protocol conversion quality evaluation includes: syntax layer protocol conversion quality evaluation, semantic layer protocol conversion quality evaluation and performance indicator protocol conversion quality; The protocol template library and / or the adversarial network are updated according to the protocol conversion quality, and / or the target resources are elastically expanded or contracted.

5. The method according to claim 1, wherein The target service in the second module calls the NLP model to perform a dual-track risk analysis on the contract text solution, including: Performing text preprocessing on the contract text scheme to obtain contract terms; Detecting specific risks corresponding to the contract terms based on a pre-set rule base; Use deep learning models to predict the risks of the contract terms and obtain potential risks; The explicit risks and the potential risks are combined to obtain risk information of the contract text solution.

6. The method according to claim 1, characterized in that The second module triggers a rolling horizon optimization algorithm based on the risk information to dynamically adjust the target solution, including: When the risk information is a high-risk clause or meets the dynamic adjustment conditions, the target solution is dynamically adjusted by adopting key parameter clipping, constraint relaxation, and parallel distributed solution to obtain the adjusted target solution.

7. The method according to claim 6, characterized in that The dynamic adjustment conditions include: key resource price fluctuations greater than a threshold, suppliers being blacklisted, and task progress deviations greater than a preset threshold.

8. An enterprise project management device, characterized in that: Applied to an enterprise project management system, the device includes: A publishing and distribution unit, configured to publish a business event to the first module of the enterprise project management system and distribute the business event to the second module via a unified data bus; A determination unit, configured for the target service in the second module to call a target model to determine a target solution corresponding to the business event; a protocol conversion unit configured to cause the second module to perform protocol conversion on the target solution using a protocol adapter, and to send the target solution after protocol conversion to an external system, so as to obtain, through the protocol adapter, a contract text solution returned by the external system for the target solution after protocol conversion; a dual-track risk analysis unit, configured to cause the target service in the second module to invoke an NLP model to perform a dual-track risk analysis on the contract text scheme to obtain risk information of the contract text scheme, wherein the dual-track risk analysis includes: a risk analysis method based on a preset rule base and a risk analysis method using a deep learning model; A dynamic adjustment unit is used for the second module to trigger a rolling time domain optimization algorithm based on the risk information to dynamically adjust the target plan to obtain an adjusted target plan.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to execute the method according to any one of claims 1 to 7.

Images