Design Of Complex Logic And Computational Reasoning Engine And Its Application In Function

In recent years,with the rapid development of artificial intelligence technology,many experts and scholars began to focus on the combination of artificial intelligence technology and education.Through the artificial intelligence technology and Internet thinking to define a new education model,to achieve computer automatic solution,automatic counseling and other functions,to provide students with fast,fair,systematic and tailored educational resources.This is of great significance to improve the quality of education,reduce the cost of education and innovative education model.The main research direction of this thesis is automatic solution based on the input of the original primary mathematics topics,referred to as "college entrance examination robot".This is the basis for all subsequent automated counseling functions,It mainly contains the following points:1)Research and construction of complex logic and computational inference engineThis thesis mainly studies the organizational structure and core algorithm design of complex logic and computing engine,and gives the corresponding solution for the emergence problem.It analyzes the structure and composition of complex logic and computational inference engine from the aspects of complex logic reasoning engine research and construction,complex logic and computational reasoning,and humanoid solving.Which focuses on the engine to build some of the three different complex logical reasoning organization,and the use of "first forward last backward" reasoning way to build reasoning engine.Then the relationship between the reasoning engine and the symbolic computing platform is also studied,and the computational service provided by the symbolic computation is supported by the computational reasoning of the concrete problem.Based on the reasoning,we design the search algorithm based on DFS to reconstruct the humanoid solving process.2)Knowledge representation of elementary mathematical function problemsWith the support of Nature Language Process(NLP),this thesis mainly studies the knowledge modeling based on first-order predicate logic for function problems in elementary mathematics.Through statistics,collation and analysis,the knowledge of the knowledge can be divided into three categories: entity knowledge,entity attribute knowledge,the knowledge between entity and entity.3)Reasoning rules for elementary mathematical functionsThe algebraic knowledge reasoning,which is represented by the function problem,has the characteristics which are completely different from the geometric knowledge reasoning: computationality and openness.Computational means that the reasoning is reflected in the calculation results,is a typical embodiment of "counting for reasoning".Openness refers to reasoning that knowledge does not converge and is an infinite set of expansion.Based on this feature,we use the combination of rule flow reasoning and general rule reasoning in rule reasoning of function problem.The rule flow allows the user to customize the process structure to match the execution order of the rules;and the general rule can perform the RHS only if the LHS condition satisfies the match.In the study of general rules,the rules are divided into calculation rules,logical rules and association rules.The rule flow allows the user to customize the structure of a similar flow chart to specify the order in which the rules match the execution.The general rule can trigger the execution of the RHS only if the LHS condition satisfies the match.In the study of the general rule part,the rules are divided into calculation rules,logical rules and association rules.It is very important to construct a complete and efficient reasoning rule base by rules and detailed rules.In this thesis,we put forward the organizational structure of the reasoning engine and the related theoretical methods,and finally construct a common complex logic and computational reasoning engine,and achieve a more complete function reasoning rule base.Which makes the function problem solving rate of 67%,and the average solution time is less than 5 minutes.