Research On Intelligent Test Paper Generation Based On Genetic Algorithm

With the rapid spread of network technology and the continuous progress of education reform, the research of intelligent test paper generation has become more and more significant to scholars and experts. The test paper generation is a constrained multi-object optimization problem and the generation method is the key factor in test paper generation efficiency and quality.This paper has described a new mathematical model of test matrix and designed a new genetic algorithm. This paper has described a new mathematical model of test matrix and designed a new genetic algorithm coding used for the storage of the paper’s properties and fitness. And the question number maps gene directly. In this paper the genetic algorithm has been improved. And improved algorithm adopts segment real number code. With the usage of a greater proportion of crossover and mutation, a kind of method of generating test initial population has been proposed in this paper. The method which takes the knowledge as the basis can meet the three basic requirement:distribution, knowledge points, answer time etc. And the paper analyzed main functions of the examination system, and a intelligent test paper generation system is designed.The results of the experimental show that the paper can be easily extended index and it can improve the operation efficiency by avoiding the repeated calculation of fitness value and decoding. Practical application results show that the proposed algorithm improves the evolutionary origin and can keep the diversity of population. In addition crossover and mutation operation conduct in the interior of each section, it may guarantee the quantity of each type correct matching and different knowledge point of the same type in the process of test paper making. Experiment result show that the new method is more effective and efficient in solving the problem of intelligent test paper generation.