Research On The Key Technology Of Performance Modeling,Resource Allocation And Optimization Of Networked Intelligent Service System

With the development and popularization of the Internet and mobile terminal technologies,the traditional business management-oriented and functional module of the management information system and its business model have undergone tremendous changes,showing three characteristics of networking,processing,and intelligence.Networking means that business can be handled with the Internet at any time anywhere by anybody.Processing means that the management information system is driven by the process.Each business is related and dependent and all businesses are flowed by processes.Intelligence means that the whole system appears to be an intelligent service system(Agent)managing all kinds of complex requests on time.We claim the application system with the characteristics of networking,processing and intelligence is the networked intelligent service system.Large e-commerce platforms represented by Taobao and large-scale complex government service platforms represented by social security are typical applications of networked intelligent service system.However,with the requirements for application of the network and increases for service efficiency,the networked intelligent service system mainly faces two challenges.One is massive requests from users.Because each service terminal of the system is open and faces to the public,more and more business can be managed through a mobile terminal by individual and that leads to massive requests and complex system load changes.The other challenge is service high efficiency.People's expectation of the service efficiency increases higher and higher that needs the system response to the request on time and to handle the service request more efficiently.This thesis aims at the two challenges of the networked intelligent service system.Aim at improving throughput ability and service efficiency of the networked intelligent service system.Focus on the modeling,performance modeling evaluation and task allocation of the networked intelligent system.The main work is summarized as follows.(1)Modeling of the net-worked intelligent service system.This thesis models the networked intelligent service system into an Agent system.Aiming at the two challenges of the networked intelligent service system,the Agent system major has two abilities that are solving the massive requests from the users and improving the service efficiency of the system.The Agent system has four elements that are environment,state,action and reward.The environment mainly refers to the user's request.The Agent system makes a decision on the task and resource allocation by perceiving the environment.Increasing service efficiency and optimization all serve as an incentive and reward for the feedback of the Agent system environment.(2)Aiming at the massive requests challenge of the networked intelligent service system,the Agent system needs the ability of solving the massive requests quickly on time.On this purpose,this thesis models the system based on the queueing theory,analyzes the performance of the networked intelligent service system,finds the bottleneck of the system,and optimizes the resource allocation of each post.(3)Aiming at the improvement service efficiency challenge of the networked intelligent service system,the Agent system needs the ability to allocate tasks scientifically and highly efficiently allocates the appropriate tasks for appropriate people or posts.On this purpose,this thesis uses the thought of crowdsourcing task allocation method and develops a networked intelligent service system task allocation algorithm.This algorithm first develops the individual task list and then allocates the resource.The research results of this thesis are applied in the medical insurance refund process.By using the modeling method,performance analysis method,optimization method and etc.to the data of the medical insurance refund process,this thesis verifies experimentally.Experiments have shown that applying the results to the medical insurance refund process improves the management efficiency of the entire medical insurance application system and the average finishing time of all the business is decreasing 59%.The research methods,analysis methods and optimization methods which are developed in this thesis for networked intelligent service systems are universal,and have great significance for use when handling other field networked service systems.