基于web考试系统的智能组卷算法的研究与实现

Research and Implementation of Algorithm of Intelligent Test-Paper Based on Web Examination System

衡量一份试卷质量高低的主要因素是试题的覆盖范围是否合理,难度是否适中,区分度是否最大等等。随着计算机技术的普及,越来越多的学校采用智能组卷系统代替原有的教师手工组卷,实现了教考分离,因此,智能组卷成功率也成为大家关注的焦点。介绍了目前流行的组卷算法,分析了各种算法的特点以及不足之处,结合组卷问题的数学模型提出了分层次的遗传算法,该算法提高了原遗传算法的局部收敛度和全局搜索性。

The main factors effecting paper quality are whether the test covers all knowledge scope, whether the difficulty of test is moderate and whether questions have largest area indexing, etc. Since most schools use intelligent test-paper system instead of the manual work with the popularization of com- puter technology, success rate of intelligent test-paper has become the focus of attention. This paper de- scribed the current popular algorithms on intelligent test-paper and analysed its features and shortcomings. At last, the author proposed a stratified genetic algorithm based on mathematic models of test paper to im- prove the local convergence of the original genetic algorithm and global search rate.