中文摘要：

遗传算法能解决选题的盲目性，并能从群体中选择更满足条件的个体，具有很强的智能性，同时它能根据不同的环境产生不同的后代，具有动态性、自适应性，从而能满足题库不断变化的要求。由于理想题库容量太，覆盖面广，因此选题计算量大。利用遗传的内在并行性，可以有效地解决计算量大的问题。而且其搜索是从一个初始种群出发，即从多点出发。减少了陷入局部最优解的概率。遗传算法优化求解只需知道问题本身所具有的目标函数，同传统优化算法比较．它具有更强的鲁棒性。本文的大量试验亦证明了这一点。

英文摘要：

Genetic Algorithms can solve the blindness of select questions, selecting better individuals which satisfy the requirement from the population. They have strong intelligence, produce different offspring according to different environment, and have dynamic, automatic adaptability which satisfies the requirement of variation of test paper databases. As the capacity of test paper database is very large, and the content is very extensive,the computation is very complicated. We can solve the complicated computation efficiently by genetic parallel property. Its search starts from an initialized population, that＇s to say starting from many points, which induces the probability of being in local optimized solutions. We just know the objective function of the problem to get optimized solution by genetic algorithms. It is more efficient compared to traditional optimized algorithms.