The Application And Research Of The Paper-Composed System Based On Genetic Algorithm

At present, with the development of the technology of computer and database, the testing system on computer gradually becomes a hotspot for research. Among this, the auto-composing test paper module decides the quality of the examination question chosen and decides if the examination system can evaluate the real level of the students and the teaching quality of the teachers. As a result, it mostly decides the quality of the examination system. Thus, the present usual auto-composing algorithm gradually represents the low success efficiency and the long time it takes, the paper composed can not satisfy the practical needs and so on. So, the purpose of this paper is to search for a more suitable and efficient auto-composing algorithm in order to satisfy the needs of the teaching work.Genetic algorithm is a high efficient and global parallel searching algorithm. It is simple and in common use and suitable for parallel deal, robustness and so on. But it exists some problems such as premature. So the paper does some research on the auto-composing problem based on genetic algorithm.The thesis firstly studies the current research status of the intelligent composing test paper and genetic algorithm, elaborates the basic theory of genetic algorithm and the defects of genetic algorithm itself and analyzes the cause of the premature. And then it elaborates the basic theory of composing system, analyses several important targets of the item bank and establishes the composing test paper module and the target function. The target function synthesizes the type, the difficulty, the discrimination and the acknowledge parameter of the item bank, which almost satisfies the needs of the user.Secondly, the problem of composing test paper is designed and the item bank is built based on the mathematic module. An algorithm of composing test paper based on dynamic auto-adaptive technology is proposed because of the character of the composing test paper and the defects of the genetic algorithm itself. Improvements are done in the encode of the chromosome, the genetic operator, the design of the controlling parameters and so on. Then the improved algorithm is used into the composing testing paper. Experiments show that the improved algorithm has a great efficiency in solving the problem of composing test paper and it has a good practicability.Finally, the thesis analyzes and designs the composing test paper system. Development of the system is done under the environment of J2EE. In the end, the thesis summarizes the full text and proposes the direction for further research.