Design And Implementation Of Plane Geometry Reasoning System Based On The Rule Engine

Artificial intelligence is a branch of computer science, which can be used to simulate and extend human intelligence as the theoretical and technical method.Consisted of natural language processing, knowledge representation, machine learning,knowledge acquisition, knowledge reasoning, perception and planning, artificial intelligence has become the subject with ample content since it was proposed in the1950 s. In the development and application of artificial intelligence, people are trying their best to build intelligent systems to simulate and the expansion of human intelligence. Based on the analysis of the expert system and the production system, we regarded the Drools rules engine as reasoning machine, used Java multi-thread technology, designed and implemented a multi reasoning machine and parallel inference system for plane geometry reasoning. This thesis mainly has carried on the following several aspects of research:Firstly, we classified the knowledge in plane geometry and selected the appropriate method of knowledge representation. In this thesis, geometry knowledge roughly divided into two categories: one is theory-based knowledge constituted by the mathematical theory of plane geometry; another is fact-based knowledge composed by objective concepts or facts. Different ways of knowledge representation was often adopted to represent these two types of knowledge. In this thesis, production rules and first-order predicate logic were used to represent theory-based and fact-based knowledges.Secondly, the principle of expert system and production system were analyzed.Expert system is an important branch of early artificial intelligence and a kind of intelligent computer systems with specialized knowledge and experience. Expert system is knowledge-based systems, using knowledge representation and knowledge reasoning technology to simulate the domain experts thinking process and solve the complex problems that only can be solved by domain experts. Production system is a computer system proposed by E.Post in 1943, which could be used as a problem expression model in artificial intelligence and the knowledge system. The basic principle of expert system and production system was briefly analyzed in this thesis.Thirdly, the Drools rules engine was used to design the basic structure of the planegeometry reasoning system. Using Drools rules engine as reasoning system’s reasoning machine, rule-based production system structure and forward reasoning approach were applied in the design of the knowledge reasoning. In this thesis, the main structure modules of the plane geometry reasoning system consisted of the system interface,controller, the Drools rules engine, rule base, fact base, graphical information and Knowledge-derived network.Finally, according to the basic structure of plane geometry reasoning system, the plane geometry inference system was implemented on the Java platform. Using Java multi-thread technology, Plane geometry reasoning was divided into geometric reasoning, algebraic reasoning and hybrid reasoning in this thesis. Therefore, in this thesis, the final plane geometry reasoning system was a parallel reasoning system which contained multiple reasoning machine and multiple knowledge base.A kind of Artificial intelligence solution to geometry reasoning was raised in this thesis, which provided some problem-solving ideas and technical methods for knowledge reasoning. Further more, based on the plane geometry reasoning system designed in this paper, a more perfect and powerful problem solving system can be to built.