Research On The Automation Of Programming Based On Expert Systems And Efficient Algorithms

Algorithmic programming automation is a technique that uses computer programs to automate the generation of algorithms.It is mainly used to solve bottlenecks in programming and to improve the efficiency and reliability of programming by automating the generation of algorithms.As the field of artificial intelligence continues to grow,it has been extensively researched and expanded in the field of computer science,however,algorithmic programming still has some challenges and limitations.Firstly,it requires a lot of time and effort.Secondly algorithmic programming needs to deal with a large amount of data and complex problems,and human negligence and errors can also lead to a certain error rate in programming.Therefore,seeking an efficient and reliable algorithmic method is an urgent need to automate programming and is also the research goal of this dissertation.In this dissertation,based on the PAR method developed by our research team under the continuous funding of several national projects,we propose a theory and framework for efficient algorithmic programming for problems with conventional solutions,and integrate it with the techniques of expert systems to realize the automatic construction of algorithmic programs;we choose Radl as the target language,develop a Radl protocol to Radl algorithmic program automatic generation system,and The reliability of the system is analyzed theoretically.Finally,the automatic conversion capability of its support tool RSPEto RALG generation system is verified by several cases.The results show that the method proposed in this paper has some research significance and can provide some reference basis for program developers.The main innovations of this dissertation are as follows:(1)A new efficient theoretical approach to algorithmic programming and a generalized algorithmic structure of problems based on the PAR method is proposed for problems with conventional solutions commonly found in programming.The knowledge of the programming process in the PAR method is classified,extracted and managed based on the technique of knowledge representation,and the Radl statute program analysis rules and Radl algorithm program generation rules are refined,which are transformed into a computer understandable form to construct a knowledge base of algorithmic programs.(2)Based on the proposed efficient algorithm theory combined with knowledge representation and knowledge inference techniques,a Radl algorithm program automation system RSPEto RALG is designed for the goal of precise intelligence to realize the automatic conversion of Radl statute to Radl algorithm program,and the system is tested and evaluated,which greatly improves the program development efficiency and the automation degree of PAR platform.