Research On Cooperative Task Decomposition Based On Knowledge Reasoning

With the popularity of computer, data generated on the Internet increases gradually, so problems needed to deal with are becoming more and more complex. Although the performance of computer hardware and software are constantly improving, but computer need dealing with complex issues, which will take a long time. In a distributed network environment, a number of complex issues may needed more than one computer together to resolve. To solve these problems, we must first break down complex issues. Task resolve is a process to make complex issues becoming simply. Reasonable resolve can greatly reduce the complexity of the task. The sets of sub-task resolved require multiple computers synergistic to complete.Collaborative control system based on multi-host has become a hot research. Simple collaboration control system can not automatically and intelligently resolve tasks. In order to make intelligent collaborative strategy automaticlly, the paper puts the knowledge base into cooperative control system. Using expert knowledge and effective knowledge reasoning resolves task for collaboration control system. It can solve complex problems effectively.With the continued expansion of the knowledge in knowledge base, some of the uncertain complex knowledge can not be showed by a simple rule. This uncertainty is mostly caused by the ambiguity, which is difficultly to be dealt with traditional two-valued logic accurately. The fuzzy technology gives the theory and method of such a vague expression of information. It realize that the natural language information transform into quantitative representation of the information. For example, using fuzzy sets can make the noumenon extended to imprecise noumenon, thus achieve representation of imprecise knowledge. Therefore, this paper combines the fuzzy technology with knowledge base to form fuzzy knowledge inference structure,which can deal with uncertainty information.To be able to solve complex problems, this paper studies the following aspects:(1) Research knowledge representation and reasoning of knowledge base, knowledge representation methods are summarized:logical representation, production representation, frame representation, object-oriented representation, web semantic representation and ontology representation. Researches the strategies and mechanisms of knowledge reasoning; Analyzes the advantages and disadvantages of forward reasoning and backward reasoning; Designes reasonable strategies to fit in this paper. Research on fuzzy logic, analyzes fuzzy reasoning model and summarize the appropriate inference method. Focus on rule-based knowledge representation, combine fuzzy logic and knowledge reasoning, proposes representation of fuzzy rules based on the credibility, and gives fuzzy inference method based on simple rules and complex rules.(2) Study the task resolve algorithm, construct the task model, and propose a<TN,ID,TG,IN,O,ST,ET> task model of seven tuples. According to the principle of task resolve, task resolve of this paper are in accordance with the principle of independence, a hierarchical principle, reducing principle, the principle of proportionality and the principle of termination. This paper considers the characteristics of the task performer, combining features of the task, propose task decomposition algorithm based on intelligent agent, and give the corresponding task decomposition process and pseudo-code description.(3) Research CLIPS expert system development tool, design the overall framework of based on CLIPS knowledge base systems, and give each module division. Extend CLIPS function, combine the fuzzy inference technology with the knowledge base system, make knowledge base with the ability of fuzzy reasoning. Design a friendly interface of knowledge acquisition, including obtaining facts, rules and define the template interface. Meanwhile, designing a fuzzy inference interface, which can present the reasoning process and results preferably. Achieve key data structures and interfaces of task resolve knowledge, and in software development tasks for instance, carry out the task resolve. Resolve the complex software development tasks into a number of simple tasks successfully. Make the execution of tasks be clear at a glance, and the division of staff is also more clearly.