Research And Implementation Of The Automatic Solving Of Inequalities In A Humanoid Answering System

With the continuous development of information technology,the development of the field of artificial intelligence is growing by leaps and bounds,in all areas of society cannot do without computer and artificial intelligence,such as education,the study of solving math problems by machine has lasted for decades,some good progress has been made in many aspects of the machine proving,depends on previous work experience and achievements,the paper will create a elementary mathematics humanoid answering system that used inequality machine solving and reasoning technology,as much as possible to solve the problems of inequality related elementary mathematics.The main work of this paper is as follows:1.Some algorithms for solving problems related to inequality are studied,improved and implemented.Using and improving Monte Carlo tree search method for inequality scaling and expression transformation,which improves the efficiency of inequality scaling and expression transformation.In the Monte Carlo tree search expression transformation method,we implement an expression deformable method based on syntax tree and parameter space,so that the expression transformation is efficient.The method of differential substitution is implemented to determine the nonnegativity of the expression.We use Grobner Basis to generate as many equations as possible for knowledge base,which not only simplifies the relationship among equations,but also can easily get the objective analytic formula and objective equality,so that we can solve the most valuable and parameter ranges problems easily.2.A set of knowledge representation method for inequality related problems is designed.The first condition of reasoning engine is to have a set of knowledge representation.After establishing appropriate knowledge representation method,the efficiency and accuracy of the system for reasoning can ensured.On the basis of this set of representations,Natural Language Processing can identify the subject,and the reasoning engine will use these relationships and entities for reasoning.3.This paper designs a set of mathematical problem reasoning engine,which includes four main engines: forward reasoning,branch reasoning,auxiliary reasoning and reverse reasoning.All reasoning works in this range.Forward reasoning is mainly responsible for general computational reasoning and logical reasoning.Branch reasoning is mainly responsible for reasoning about the branches of different parameter ranges of parameter discussions and multiple choice.The auxiliary reasoning is mainly based on intuitionistic reasoning and computational reasoning.The reverse reasoning is start from the conclusion and based on logical reasoning,computational reasoning and intuitionistic reasoning.In addition,a set of inequalities rule base is created and maintained,and reasoning rules are the basis of reasoning engine.The four engines run on the basis of these reasoning rules.4.The ideas and methods of solving the classical problems in inequality related problems in elementary mathematics are given and realized,thus a set of automatic reasoning system for inequality problems is realized.At the same time,a multi solution technique is also implemented to help the system to enhance its effectiveness.Lastly,for the last step of solving the inequality problem,the problem of halting,the corresponding solution is given so that the system can be stopped as accurately as possible.After a series of system building and adding,perfecting and testing methods,the correct rate of solving the intermediate difficulty inequality problems in the elementary mathematics for the humanoid answering system is 73.25%.