基于改进遗传算法的组卷策略研究  

Intelligent Test Paper Auto-Generation Based on Improved Genetic Algorithm

作　　者：唐金文[1] 

机构地区：[1]曲靖师范学院软件工程研究所,云南曲靖655011

出　　处：《曲靖师范学院学报》2014年第6期27-31,共5页Journal of Qujing Normal University

摘　　要：针对题库系统的自动组卷策略,采用分段整数编码、初始种群满足试卷部分指标,其余部分指标预设权重的改进遗传算法.在预设权重指标中,引入知识点分数分布或知识点分数分布百分比指标代替传统的覆盖率指标,结合试卷的难度系数和区分度系数指标,设计一个较为合理的个体适应度函数来评价个体;在遗传进化过程中,引入个体相似度指标来评价种群中个体的雷同程度,并给出个体相似度计算方法.For the problem of automatic test paper, genetic algorithm is improved by paragraphed integer cod- ing and initial populations satisfy part of indicator for the test paper, and other indicator preset weights. A more reasonable individual fitness function to evaluate the individual is designed in the preset weight indica- tors, percentage of knowledge score distribution or knowledge score distribution instead of the traditional cov- erage indicators combined with the difficulty of the examination paper coefficient and degree of differentiation coefficient indicators. In the process of genetic evolution, the individual similarity indicators are introduced to evaluate the individual＇s similar degree, and the similar calculation method is given.

关 键 词：自动组卷 遗传算法 分数分布 个体相似度 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]