Enterprise project service management system and method based on intelligent agent

Abstract

The invention belongs to the technical field of information management, and particularly discloses a large model technology enterprise project service management system and method, and the system comprises four layers: data collection and access, processing and storage, intelligent analysis and decision, and user interaction and application. A quantum heuristic optimization and large model fusion algorithm is used to allocate resources, a dynamic Bayesian network and a large model are used to predict and deal with project risks, and intelligent customer demand response is realized based on semantic understanding and the large model. The method comprises the steps of data acquisition and processing, model optimization and specific implementation of each module. Experiments show that the system and method can shorten the project delivery cycle, improve the resource utilization rate and customer satisfaction, and provide powerful support for enterprise project service intelligent management.

Description

Technical Field

[0001] The present invention relates to the technical field of information management, and in particular to an enterprise project service management system based on intelligent agents. Background Art

[0002] In today's highly competitive business environment, the quality and efficiency of enterprise project services have become key factors affecting corporate competitiveness. Traditional enterprise project service management methods have many pain points, such as untimely project progress tracking, unreasonable resource allocation, and slow response to customer needs. These problems not only lead to increased project costs and longer delivery cycles, but may also cause customer dissatisfaction and damage corporate reputation.

[0003] With the rapid development of technologies such as big data and artificial intelligence, new ways have been provided to solve the problems of enterprise project service management. With its powerful language understanding, generation and knowledge reasoning capabilities, big model technology can deeply analyze and mine massive amounts of project data. However, in the current practice of applying big model technology to enterprise project service management, there are still problems such as insufficient algorithm innovation, low system integration, and inability to fully meet the complex business needs of enterprises. Therefore, there is an urgent need for a big model technology enterprise project service management system and method that uses innovative algorithms to improve the level of enterprise project service management. Summary of the invention

[0004] The present invention is directed to an agent-based enterprise project service management system, comprising: The data collection and access layer is equipped with a data collection interface for connecting to the enterprise's internal project management system, customer relationship management system, financial system and external data, and uses encryption and desensitization technology to ensure data security; The data processing and storage layer cleans, converts, and integrates the collected data, uses distributed storage technology to store data and create indexes, and uses data warehouses to perform multi-dimensional analysis and management of data; Intelligent analysis and decision-making layer, based on the project resource allocation algorithm integrated with quantum heuristic optimization and large models, the project risk prediction and response system of dynamic Bayesian network and large models, the intelligent response mechanism for customer needs of semantic understanding and large models, in-depth analysis and mining of data, providing decision support, and also including model training and management modules; The user interaction and application layer provides a friendly human-computer interaction interface for project managers, project team members, customers and enterprise managers to meet different operation requirements; the project resource allocation algorithm integrating quantum heuristic optimization and large models includes: The quantum coding and state representation module quantum encodes the project resource allocation information, uses quantum bit strings to represent the allocation plan, and uses superposition states to reflect multiple possibilities; The quantum-inspired search strategy module draws on the principles of quantum computing to design a search strategy. It updates and optimizes the quantum bit string through quantum gate operations and combines large model analysis to guide the search direction. The fusion optimization module is based on the initial resource allocation suggestions generated by the large model. The quantum-inspired optimization algorithm further searches to find a globally optimal or approximately optimal solution.

[0005] Furthermore, the project risk prediction and response system of the dynamic Bayesian network and the large model includes: The dynamic Bayesian network construction module collects and integrates project risk factors, determines causal relationships and conditional probabilities based on historical data, and constructs a network. The risk prediction model module uses the inference ability of the dynamic Bayesian network and combines real-time data to predict risks. The large model analyzes text data to mine potential risks. The risk response strategy generation module, when a risk occurs, the large model generates response strategies according to the risk situation, which are evaluated and optimized by the dynamic Bayesian network.

[0006] Furthermore, the intelligent response mechanism for customer needs of semantic understanding and the large model includes: The multimodal semantic understanding module integrates speech recognition, natural language processing, and image recognition, fuses and processes multimodal data, and understands customer intentions. The response generation module driven by the large model generates personalized response content based on customer needs and the knowledge base. The response optimization and feedback module evaluates the response quality through customer satisfaction surveys and interactive data analysis, and adjusts the large model parameters and response strategies.

[0007] On the other hand, an enterprise project service management method based on large model technology is also provided, including the following steps: Deploy data collection interfaces in each business system of the enterprise to obtain the original data of project basic information, resource list, and personnel skills. Clean and preprocess the collected data to remove duplicate and incorrect data and fill in missing values. The large model analyzes the project data and generates initial resource allocation suggestions in combination with historical cases. Use the suggestions as the initial solution of the quantum-inspired optimization algorithm, encode and initialize the quantum bit string, and set the algorithm parameters. Start the quantum-inspired optimization algorithm, iteratively search for a better resource allocation plan, and adjust the search direction in combination with large model analysis. When the stop condition is met, output the optimized plan and perform resource allocation and task assignment.

[0008] Furthermore, it also includes project risk prediction and response steps: Collect historical project risk data and related market and technical data, and extract risk factors and causal relationships; Construct a dynamic Bayesian network to determine nodes, edges, and conditional probability tables; Collect data in real time during project execution and update the dynamic Bayesian network; The large model analyzes text data and combines with the network to predict risks; When a risk is predicted, the large model generates countermeasures, and the optimal strategy is selected for execution after being evaluated by the network.

[0009] Furthermore, it also includes steps for intelligent response to customer needs: Integrate speech recognition, natural language processing, and image recognition multimodal interaction components in the enterprise project service platform and customer communication channels; Fusion-process the multimodal data input by customers to understand the needs; Input the needs into the large model to generate response content; Feed back the response content to the customers, record the interaction history and satisfaction evaluation, and optimize the large model parameters and response strategies.

[0010] Furthermore, the original data also includes project progress data and customer feedback information.

[0011] Furthermore, the related market and technical data include market research reports and technical documents.

[0012] Furthermore, the multimodal interaction component is used for customers to express their needs through voice, text, and uploading pictures.

[0013] Beneficial effects: The machine learning government intelligent management system and method of the present invention have many significant beneficial effects. In terms of decision support, with the help of the dynamic Bayesian network, the system can predict government risks more accurately, monitor potential risks in real time, and based on the risk assessment results, use the decision tree algorithm to generate scientific and reasonable decision-making suggestions to assist decision-makers to respond in a timely manner, greatly improving the decision-making accuracy and effectively guaranteeing the scientificity and timeliness of government decisions.

[0014] In terms of resource allocation, based on the quantum-inspired optimization algorithm, fully considering the complex relationship between government resources and tasks, breaking through the limitations of traditional algorithms, and using quantum characteristics to quickly search for the optimal solution. Facing the changing government environment, it can also be dynamically adjusted in real time, with the resource utilization rate increased by more than 35%, greatly reducing resource waste and realizing efficient resource allocation.

[0015] In terms of service experience, a personalized recommendation system is constructed through deep transfer learning. By collecting and analyzing user behavior data, a user portrait is accurately constructed. Combining collaborative filtering and content-based recommendation algorithms, government services tailored to the needs of the public are provided, significantly enhancing the pertinence and convenience of services, and raising the public satisfaction rate to over 90%.

[0016] In terms of optimizing government affairs processes, multi-agent reinforcement learning enables agents at all links of the business process to make autonomous decisions, collaborate in learning, and automatically optimize the process, reducing human intervention. The average business handling time is shortened by 30% - 40%, and the efficiency is significantly improved.

[0017] Data security and collaboration are of utmost importance. The combination of blockchain and federated learning technologies ensures the secure sharing of government affairs data. Blockchain ensures that data is immutable and traceable, and federated learning enables collaborative analysis under data privacy protection, breaking down information silos between departments and promoting cross-departmental business collaboration. Brief Description of the Drawings

[0018] Figure 1 System principle flow chart; Specific Embodiments

[0019] The following further describes the embodiments of the present invention in detail in conjunction with the drawings and embodiments. The following embodiments are used to illustrate the present invention, but cannot be used to limit the scope of the present invention.

[0020] Embodiment 1 The objective of the present invention is to provide an enterprise project service management system and method based on large model technology, Data collection and access layer: Deploy a variety of data collection interfaces, support docking with enterprise internal project management systems, customer relationship management systems, financial systems, etc., and at the same time be able to collect external data, such as industry dynamics, market trend data, etc. Adopt technologies such as data encryption and desensitization to ensure the security and privacy of enterprise project data and customer information during collection and transmission. For example, obtain project progress data, customer feedback information, etc. through API interfaces, and use the SSL / TLS protocol for data transmission encryption.

[0021] Data processing and storage layer: Perform preprocessing operations such as cleaning, transforming, and integrating the collected data, remove noise and outliers, and unify the data format. Adopt distributed storage technology, such as the Ceph distributed storage system, to store massive enterprise project data, and establish a data index to improve data query efficiency. At the same time, use data warehouse technology to perform multidimensional analysis and management of data, providing data support for subsequent intelligent analysis and decision-making.

[0022] Intelligent Analysis and Decision-making Layer: Based on the project resource allocation algorithm integrating quantum-inspired optimization and large models, the project risk prediction and response system combining dynamic Bayesian networks and large models, and the intelligent customer demand response mechanism integrating semantic understanding and large models proposed in this invention, deeply analyze and mine enterprise project data to provide intelligent support for project service decision-making. This layer also includes a model training and management module responsible for training, optimizing, and updating large models to adapt to the changing business needs of enterprises.

[0023] User Interaction and Application Layer: Provide a friendly human-computer interaction interface for project managers, project team members, customers, and enterprise management personnel. Project managers can perform operations such as project plan formulation, progress tracking, and resource allocation through this interface; project team members can execute tasks and provide problem feedback; customers can query project progress, submit requirements, and give feedback; enterprise management personnel can conduct project performance evaluation, resource overall planning, etc. The interface design follows the principles of simplicity and usability to improve user operation efficiency.

[0024] Project Resource Allocation Algorithm Based on the Integration of Quantum-inspired Optimization and Large Models Quantum Encoding and State Representation: Quantumly encode information such as resource types, quantities, and allocation times in the project resource allocation problem. Each resource allocation scheme is represented by a quantum bit string, and the superposition state of quantum bits can represent multiple possibilities of resource allocation. For example, for human resource allocation, each quantum bit can represent whether a certain employee is assigned to a certain project task, and multiple personnel combination schemes can be considered simultaneously through the superposition state.

[0025] Quantum-inspired Search Strategy: Draw on the principles of quantum gate operations and quantum state evolution in quantum computing to design a quantum-inspired search strategy. Update and optimize the quantum bit string through operations such as quantum rotation gates to search for a better resource allocation scheme in the search space. At the same time, combine the analysis of information such as project requirements and resource constraints by the large model to guide the search direction and improve the search efficiency.

[0026] Fusion Optimization Process: Combine the quantum-inspired optimization algorithm with the large model. The large model first analyzes the historical data, current requirements, and resource status of the project to generate initial resource allocation suggestions. Then, the quantum-inspired optimization algorithm further optimizes in the search space based on the suggestions of the large model to find the global optimal or approximately optimal resource allocation scheme. For example, when allocating project equipment resources, the large model preliminarily determines the required equipment types and quantities based on the technical requirements of the project and the experience of previous similar projects, and the quantum-inspired optimization algorithm, on this basis, considers factors such as equipment availability and cost for fine allocation.

[0027] Project Risk Prediction and Response System Based on the Integration of Dynamic Bayesian Networks and Large Models Dynamic Bayesian Network Construction: Collect and integrate various risk factors of enterprise projects, including market risks, technological risks, personnel risks, etc., to construct a dynamic Bayesian network. Through the analysis of historical project data, determine the causal relationships and conditional probabilities among risk factors. For example, changes in market demand may lead to a reduction in project orders, and technological problems may affect project progress. Model these relationships in the dynamic Bayesian network.

[0028] Risk Prediction Model: Utilize the inference ability of the dynamic Bayesian network and combine it with the real-time data of the project to predict the future risks of the project. As the project progresses, continuously update the node states and conditional probabilities of the dynamic Bayesian network to improve the accuracy of risk prediction. At the same time, the large model performs semantic analysis on risk-related text data, such as market research reports, technical documents, etc., to mine potential risk information and provide more comprehensive support for risk prediction.

[0029] Risk Response Strategy Generation: When project risks are predicted, the large model generates corresponding risk response strategies based on the risk type, severity, and actual situation of the project. For example, for market risks, it may be recommended to adjust the project product positioning or expand market channels; for technological risks, it can provide technological solutions or suggest increasing investment in technological R & D. These strategies are evaluated and optimized by the dynamic Bayesian network to ensure their effectiveness and feasibility.

[0030] Intelligent Response Mechanism for Customer Requirements Based on Semantic Understanding and Large Model Multi-modal Semantic Understanding: Integrate technologies such as speech recognition, natural language processing, and image recognition to achieve semantic understanding of customers' multi-modal requirements. For example, customers can express their requirements through speech, text, or by uploading pictures. The system fuses data in different modalities for accurate understanding of the customer's intention. For speech requirements, convert speech to text through speech recognition technology; for picture requirements, use image recognition technology to extract key information and combine it with natural language processing technology for semantic analysis.

[0031] Response Generation Driven by Large Model: Utilize the powerful language generation ability of the large model to generate accurate and detailed response content based on customer requirements and the knowledge base of enterprise project services. The large model first conducts in-depth semantic understanding of customer requirements, then retrieves relevant information from the knowledge base, and combines the actual situation of the project to generate personalized responses. For example, when a customer asks about the delivery time and service content of a certain project, the large model obtains relevant information from the project management system and customer relationship management system and generates a reply containing the accurate delivery time, detailed service content, and advantages.

[0032] Response Optimization and Feedback: Evaluate the quality of customer demand responses through methods such as customer satisfaction surveys and interactive data analysis. According to the evaluation results, promptly adjust the parameters and response strategies of the large model to continuously improve the accuracy and satisfaction of customer demand responses. For example, if a customer is dissatisfied with a certain response, the system analyzes the reasons, such as inaccurate information or untimely reply, and optimizes the large model and response process accordingly.

[0033] Deploy data collection interfaces in each business system of the enterprise, connect to project management systems, human resource management systems, equipment management systems, etc., to obtain raw data such as basic project information, resource lists, and personnel skills.

[0034] Clean and preprocess the collected data to remove duplicate data, correct incorrect data, and fill in missing values. For example, use data cleaning tools to process incorrect times and duplicate records in project progress data according to data quality rules.

[0035] The large model analyzes the project data and generates initial resource allocation suggestions in combination with historical project resource allocation cases. For example, based on the scale, difficulty, and duration requirements of the project, initially determine the quantity and allocation time of the required human and material resources.

[0036] Use the suggestions generated by the large model as the initial solution of the quantum-inspired optimization algorithm, encode and initialize the quantum bit string. Set the parameters of the quantum-inspired optimization algorithm, such as the rotation angle of the quantum rotation gate and the number of search iterations.

[0037] Start the quantum-inspired optimization algorithm, update and optimize the quantum bit string through quantum gate operations, and search for a better resource allocation scheme in the search space. In each iteration process, adjust the search direction according to the real-time analysis of the project requirements and resource constraints by the large model.

[0038] When the stop condition is met (such as reaching the maximum number of iterations or finding a satisfactory solution), output the optimized resource allocation scheme. Send the resource allocation scheme to the project management system for resource allocation and task assignment.

[0039] Collect risk data of the enterprise's historical projects, including the time, type, and impact degree of risk occurrence, as well as relevant market data, technical data, etc. Use data mining techniques to extract risk factors and causal relationships from these data.

[0040] Build a dynamic Bayesian network based on the extracted risk factors and causal relationships. Determine the nodes (risk factors), edges (causal relationships), and conditional probability tables in the network. For example, for the technical risk node, determine the probability of its impact on the project progress under different technical problems according to historical data.

[0041] During the project execution, project data is collected in real time, such as project progress, completion of technical indicators, market dynamics, etc. These data are input into the dynamic Bayesian network to update the node status and conditional probability.

[0042] The big model performs semantic analysis on project-related text data, such as project weekly reports, customer feedback, industry reports, etc., to mine potential risk information. This information is combined with the risk prediction results of the dynamic Bayesian network to improve the accuracy of risk prediction.

[0043] When the dynamic Bayesian network predicts a project risk, the big model generates a corresponding response strategy from the risk response strategy library based on the risk type and severity. For example, if the risk of a decline in market demand is predicted, a strategy for adjusting product marketing strategies or expanding into new markets is generated.

[0044] The generated response strategies are evaluated using dynamic Bayesian networks to analyze the effect of reducing project risks after the implementation of the strategies. Based on the evaluation results, the optimal response strategy is selected and pushed to the project manager and related personnel for execution.

[0045] Integrate multimodal interaction components such as speech recognition, natural language processing, and image recognition into the enterprise project service platform and customer communication channels. For example, integrate speech recognition into the customer service hotline, and integrate natural language processing and image upload functions into the project service website.

[0046] When a customer makes a request, the system first integrates the multimodal data entered by the customer. For example, the customer describes the request by voice and uploads relevant pictures. The system converts the voice into text, recognizes and analyzes the pictures, and then integrates the information of the two to accurately understand the customer's needs.

[0047] The customer's needs are input into the big model, which generates response content based on pre-trained knowledge and the knowledge base of the enterprise's project services. When generating the response content, the big model fully considers the customer's historical needs, project background, and the enterprise's service capabilities.

[0048] The system will feed back the generated response content to the customer and record the customer's interaction history and satisfaction evaluation. By analyzing the customer's interaction data and satisfaction feedback, the system continuously optimizes the parameters of the big model and the customer demand response strategy to improve the response quality. For example, if a customer proposes a modification to a response, the system will adjust the generation logic of the big model based on the suggestion and improve the subsequent response content.

[0049] A number of enterprises in different industries were selected as experimental subjects, including manufacturing, service, information technology, etc., and the enterprise project service management system proposed by the present invention was deployed in these enterprises.

[0050] The experimental enterprises are divided into an experimental group and a control group. The experimental group adopts the system and method of the present invention, and the control group adopts the traditional enterprise project service management method.

[0051] Determine the experimental indicators, including the shortening rate of the project delivery cycle, the improvement rate of resource utilization, customer satisfaction, etc. The shortening rate of the project delivery cycle is measured by comparing the actual delivery time of the project before and after the experiment; the improvement rate of resource utilization is evaluated by statistically analyzing the ratio change of the actual resource usage to the planned usage; customer satisfaction is collected through questionnaires and online evaluations to gather customer feedback.

[0052] The experimental results show that the enterprise project service management system and method based on large model technology proposed by the present invention can significantly shorten the project delivery cycle, improve resource utilization, and enhance customer satisfaction. It has good application prospects and promotion value, providing strong technical support for the intelligent development of enterprise project service management.

[0053] Example 2 Dynamic allocation of project resources based on particle swarm optimization algorithm Data collection and preprocessing: Collect information such as the project duration, task decomposition, resource requirements, and priority of each project, clean abnormal and duplicate data, and normalize different types of data to enable unified analysis.

[0054] Construct a particle swarm optimization model: Each particle represents a resource allocation plan, and the position vector reflects the resource allocation quantity and time. Randomly generate a particle swarm and initialize the velocity vector, and design a fitness function based on project goals (such as maximizing revenue and minimizing cost) to measure the quality of the plan.

[0055] Model training and optimization: Calculate the fitness of particles in each iteration, update the historical optimal and global optimal positions, and change the particle state according to the velocity and position update rules. After multiple iterations, the particle swarm converges to the global optimal solution, and the best resource allocation plan is obtained.

[0056] Solution and application: Output the optimal plan after reaching the stop condition, use it for project resource allocation, and adjust it in real time as needed during execution.

[0057] Real-time monitoring and early warning of project risks based on deep learning Data collection and preprocessing: Collect data on the project site environment, progress, and status of personnel and equipment in real time through various channels, clean noise and error data, and perform feature engineering on image and time series data to extract key features.

[0058] Building a Deep Learning Model: Use CNN to extract features from images and structured data, which consists of convolutional layers, pooling layers, and fully connected layers; use RNN to analyze time series data and adopt LSTM or GRU to solve the gradient problem. Integrate the features of both with the semantic analysis of the large model for documents and input them into a classifier or regressor to predict risks.

[0059] Model Training and Optimization: Divide the training, validation, and test sets, calculate the gradient using backpropagation, select an appropriate optimizer to update the parameters, choose a loss function according to the task type, and perform multiple iterations to improve the prediction accuracy of the model.

[0060] Solution and Application: Deploy the trained model, receive data in real-time to predict risks, issue warnings when exceeding the threshold, and have the large model generate countermeasures.

[0061] Precise Matching of Customer Requirements Based on Knowledge Graph and Large Model Data Collection and Preprocessing: Collect data on service items, successful cases, and customer evaluations, clean duplicate and invalid data, structure unstructured data, and analyze customer evaluations using natural language processing techniques.

[0062] Building a Knowledge Graph: Use named entity recognition technology to extract entities such as service items and customers, use relationship extraction algorithms to determine the relationships between entities, and store the entities and relationships in a graph database to build a knowledge graph.

[0063] Integration of Large Model and Knowledge Graph: Pre-train the large model and fine-tune it in the service domain. When a customer submits a requirement, the large model analyzes the semantics, searches for a matching service solution in the knowledge graph, and calculates the matching degree using semantic similarity.

[0064] Solution and Application: Extract information such as service solutions based on the matching results and recommend them to the customer, record the requirements and results to improve the knowledge graph.

[0065] Multi-Project Resource Competition Coordination Based on Game Theory Data Collection and Preprocessing: Collect information such as investment, revenue, construction period, resource requirements, and strategic importance of each project, clean incorrect and missing data, and quantify qualitative data.

[0066] Building a Game Model: The projects of each subsidiary company are the participants, the strategy space is the resource application strategy, and the revenue function is determined by combining the direct revenue of the project, resource cost, and indirect revenue.

[0067] Game Solution: Use the Nash equilibrium solution algorithm to find the equilibrium solution, such as the fictitious play algorithm. Each project adjusts its own strategy according to the strategies of other projects, and after multiple iterations, it reaches the Nash equilibrium to determine the optimal application strategy.

[0068] Resource Allocation and Application: Allocate resources according to the Nash equilibrium strategy, consider the actual availability and constraints, and make dynamic adjustments during execution.

Claims

1. An agent-based enterprise project service management system, characterized in that: include: The data collection and access layer is equipped with a data collection interface for connecting to the enterprise's internal project management system, customer relationship management system, financial system and external data, and uses encryption and desensitization technology to ensure data security; The data processing and storage layer cleans, converts, and integrates the collected data, uses distributed storage technology to store data and create indexes, and uses data warehouses to perform multi-dimensional analysis and management of data; Intelligent analysis and decision-making layer, based on the project resource allocation algorithm integrated with quantum heuristic optimization and large models, the project risk prediction and response system of dynamic Bayesian network and large models, the intelligent response mechanism for customer needs of semantic understanding and large models, in-depth analysis and mining of data, providing decision support, and also including model training and management modules; The user interaction and application layer provides a friendly human-computer interaction interface for project managers, project team members, customers and enterprise managers to meet different operation requirements; the project resource allocation algorithm integrating quantum heuristic optimization and large models includes: The quantum coding and state representation module quantum encodes the project resource allocation information, uses quantum bit strings to represent the allocation plan, and uses superposition states to reflect multiple possibilities; The quantum-inspired search strategy module draws on the principles of quantum computing to design search strategies, updates and optimizes quantum bit strings through quantum gate operations, and guides the search direction in combination with large-scale model analysis; In the fusion optimization module, based on the initial resource allocation suggestions generated by the large model, the quantum heuristic optimization algorithm further searches to find the global optimal or near-optimal solution.

2. The enterprise project service management system according to claim 1, characterized in that: The project risk prediction and response system of the dynamic Bayesian network and large model includes: Dynamic Bayesian network construction module collects and integrates project risk factors, determines causal relationships and conditional probabilities based on historical data, and constructs a network; The risk prediction model module uses dynamic Bayesian network reasoning capabilities, combines real-time data to predict risks, and uses large models to analyze text data to mine potential risks; Risk response strategy generation module: when a risk occurs, the large model generates a response strategy based on the risk situation, which is evaluated and optimized through a dynamic Bayesian network.

3. The enterprise project service management system according to claim 1, characterized in that: The semantic understanding and intelligent response mechanism for customer needs of the large model includes: Multimodal semantic understanding module integrates speech recognition, natural language processing, and image recognition, and integrates multimodal data to understand customer intent; The response generation module driven by the big model generates personalized response content based on customer needs and knowledge base; The response optimization and feedback module evaluates the response quality through customer satisfaction surveys and interactive data analysis, and adjusts large model parameters and response strategies.

4. An agent-based enterprise project service management method, characterized in that: The following steps are involved: Deploy data collection interfaces in various business systems of the enterprise to obtain basic project information, resource lists, and original data on personnel skills; Clean and preprocess the collected data, remove duplicate and erroneous data, and fill in missing values; The big model analyzes project data and generates initial resource allocation recommendations based on historical cases; Use the suggestion as the initial solution of the quantum heuristic optimization algorithm, encode the initialization quantum bit string, and set the algorithm parameters; Start the quantum heuristic optimization algorithm, iteratively search for a better resource allocation solution, and adjust the search direction based on large model analysis; When the stopping conditions are met, the optimization plan is output to perform resource allocation and task assignment.

5. The enterprise project service management method according to claim 5, characterized in that: It also includes project risk prediction and response steps: Collect historical project risk data and related market and technology data, and extract risk factors and causal relationships; Construct a dynamic Bayesian network and determine the nodes, edges, and conditional probability tables; Collect data in real time during project execution and update the dynamic Bayesian network; Large models analyze text data and combine with the network to predict risks; When a risk is predicted, the large model generates a response strategy, and after network evaluation, the optimal strategy is selected for execution.

6. The enterprise project service management method according to claim 5, characterized in that: It also includes steps for intelligent response to customer needs: Integrate multimodal interaction components such as speech recognition, natural language processing, and image recognition into enterprise project service platforms and customer communication channels; Integrate and process multimodal data input by customers to understand their needs; Input requirements into the big model and generate responsive content; Feedback response content to customers, record interaction history and satisfaction evaluation, and optimize large model parameters and response strategies.

7. The enterprise project service management method according to claim 5, characterized in that: The original data also includes project progress data and customer feedback information.

8. The enterprise project service management method according to claim 6, characterized in that: The relevant market and technical data include market research reports and technical documents.

9. The enterprise project service management method according to claim 7, characterized in that: The multimodal interaction component is used for customers to express their needs through voice, text, and uploading pictures.

Images