摘要
一个自动组卷系统的组卷效率和组卷质量取决于算法和试题数据库的结构.本文针对库结构和算法进行改进,提出基于知识点约束的分题型试题库结构及分组实数编码的遗传算法,采用具有自适应性的交叉和变异遗传算子,成功解决了多目标随机搜索时易产生死区和死循环的问题,提高了搜索速度和效率.根据试题的难度和试题的时间对试题进行赋分,并以每个知识点的分值之和占试卷总分值的比例作为知识点搜索条件,加入到适应函数中,使知识点定位更加合理准确.
The efficiency and the quality of an auto-generating test paper system depend on the arithmnetic of organizing examination paper and the structure of the test question database. This article describes the improvement of structure of the database and arithmetic, and gives a new arithmetic, which bases on the knowledge points constraint, the question type, the separate group and the truthful code. The new method also uses self-adaptive cross operator and mutation operator, and it can avoid dead-band and dead-circulation as multiple target searching. It has improved researching speed and efficiency. The mark of questions is given based on the difficulty and time. The research for the knowledge points depends on the rate of each knowledge point＇s mark to all points＇ mark.The rate is added in adaptation function. And it makes more reasonable and accurate to locate knowledge points.
出处
《山东师范大学学报（自然科学版）》
CAS
2006年第3期39-42,共4页
Journal of Shandong Normal University(Natural Science)
关键词
知识点
胚库结构
遗传算法
自动组卷
实数编码
knowledge point
structure of the database
genetic algorithm
auto-generating test paper
truthful code
