一种基于CHC算法的自动组卷方法*

利用改进的遗传算法——跨世代异物种重组大变异(cross generation heterogeneous recombination cataclysmic mutation,CHC)算法提出了一种自动组卷方法。初始种群即初始试卷集利用具有启发式信息的搜索算法产生；适应度函数是用户指定的试卷总体指标与试卷实际指标绝对误差的加权和；选择操作群体为当前群体与上世代群体的群体总和，因为大个体群操作可以更好地保持遗传多样性；交叉操作采用单点交叉方法。变异操作的步骤是：从上世代个体中挑选适应度较差的个体，对其中的

Approach of auto generating examination paper based on CHC algorithm

An approach of auto generating examination paper based on the improved genetic algorithm CHC algorithm was proposed.The initial population was produced by the searching algorithms containing heuristic information.The fitness function was the weighted sum of the absolute error of the paper entire index and actual index.The select operating population was summation of current and previous one,for the operating population was big,the genetic diversity could be kept better.The single node crossover was used in cross operating.The mutate operating process was that some individuals of bad fitness were chosen from last generation first,then locas in proportion with these individuals were picked up,these place values were determined randomly.Due to the restrict of the capability,the algorithm can avoid blindness and can speed up the constringency.Simulation results show that the planning algorithm is faster than the basal genetic algorithms and meet the optimal requirements.