Research On Composition Papers System Based On Improved Genetic Algorithms

With the development of Internet and Computer Science,Internet is used in more and more field including education.There are lots of so called distance learning system.Online learning plays an important role, so it is necessary to develop an online exercise system to keep up with the developing steps of distance learning systems. With the development of network technology, the methods of test changes greatly. The computerized test system can shorten the circle of exam work flow lessen the repeating work of teachers, and improve the efficiency. The intelligent paper forming function it supports can improve the quality of test papers and increase the objectivity of test.As for the examination database system, there are two most radical functions. One is generating a test paper which is up to mustard. The other is analyzing the result of the test paper. Intelligent generating test paper can combine knowledge and generating test paper's experience of experts with technology of computer most effectively, and generates more objective and standard test paper which is up to mustard. At the same time, it is used more expediently. It also doesn't make any professional request to operators. We can get more exact situation of students'mastery of knowledge by scientific analyzing on the result of test paper which is generated by intelligent generating test paper system. According to the feedback information of the result, we can diagnose difficult knowledge of student, and feed back it to the system and improve intelligent generating test paper system. Then we can guide students to study pertinence. Lately, there are many inland units are researching and using intelligent generating test paper system and different kinds of methods of analysis on test paper, and put it over in many facts. But there are some defects in the functions and performances, particularly on the practicability and the intelligent. So researching and developing stronger practicability and higher intelligent of generating test paper in examination database system in order to help teachers and students during the working and studying to use it expediently, those need to solve the problems currently.This article focuses on improved genetic algorithm is based on systematic research and realization of test paper. Computer intelligence test paper aims to meet the user needs to solve a set of optimal parameters of paper, so the problem is a typical test paper constraint satisfaction problem solving process. The genetic Algorithm(GA) is a new kind of whole optimization algorithm, and it uses the biologic genetics'point of view to raise the individual's adapt ability by the natural selection, heredity, mutation. The basic theories and principles of test paper generation are expounded and the constraint conditions are summarized. Based on them, a mathematical model of test paper generation is established. Then for the genetic algorithm itself vulnerable premature convergence of the defect, to introduced a genetic algorithm causes of precocious puberty, common preventive measures. In the traditional genetic algorithm based on the algorithm chosen to select some of the methods of simulation of the Niche, and for the genetic manipulation of individual replacement strategy to improve. Test by comparing the algorithm to improve the global optimization performance, efficiency and effectiveness of the verification. Test shows that the improved algorithm has better ability to overcome early in the overall search performance and convergence rate than the traditional genetic algorithm has significantly improved. Through simulation experiments show that the improved algorithm based on intelligent speed Algorithm Paper. Paper quality is better able to meet the needs of the actual test papers.