基于小生境遗传算法的自动组卷

通过对当前自动组卷方法的分析,将小生境技术引入到遗传算法自动组卷中,以期望解决遗传算法组卷的早熟问题,提出一种基于小生境遗传算法的组卷方法.该方法采用功能段结构的整数编码方式,可以克服常规采用二进制编码搜索空间过大和编码长度过长的缺点,提高求解速度和精度,同时减少迭代次数加快算法收敛.在组卷模型中以题型、题量和分值为基础,在形成初始种群和进化的过程中始终保持题型、题量和分值不变,从而简化优化目标.试验结果显示,该方法能有效限制种群中相似个体的过多复制,从而维持群体的多样性,抑制出现早熟现象,改进遗传算法在自动组卷中应用的效率,其运行时间更短,误差更小.

Auto-generating test papers based on niche genetic algorithm

The niche technology was introduced into auto-generating test papers with genetic algorithm by means of analyzing the method of present auto-generating test paper in anticipation of solving the problem of prematurity of test paper with genetic algorithm,so that a test papers method was presented on the basis of niche genetic algorithm.In this method the integer coding mode with function-block structure was used to eliminate the defect of excessive searching space and code length in the conventional binary c...