交互式遗传算法中收敛性及用户评估质量的提高

交互式遗传算法（IGA）有一个明显的缺点，即用户在交互过程中，对每代的所有个体都要进行评估，当个体的数量较大，进化代数较多时，用户容易疲劳。针对这一问题本文提出了三种解决方法：第一，利用粗糙集（rough set)理论作为信息提取技术，有效地提高交互式遗传算法的收敛性，降低进化代数。第二，引进一个自主式遗传操作过程，以缓解用户在交互过程中的疲劳性。第三，利用相似距离值这个新概念提高用户对个体评估的精确度。最后在人脸图形上进行实验的结果表明，本文所提的三种方法可以有效的缓解交互式遗传算法过程中用户疲劳问题。

Improvement on Convergence and Quality of User Evaluation in Interactive Genetic Algorithms

Interactive genetic algorithm (IGA) has an obvious shorteoming: one must evaluate all individuals for each generation in the interactive process and is easily prone to fatigue when there are a large number of individuals and many evolutionary generations. To deal with this problem, three approaches are proposed in this paper. First, The rough set theory can be used as information retrieval technique. This method can effectively improve the convergence and reduce the number of evolutionary generations of IGA. Second, An autonomous GA stage is introduced. This combination can not only consider human interests in evaluation process, but also substitute human evaluation process by machine learning method under the condition of human fatigue. Third, a new concept is put forward about similar distance. It can effectively improve the precision of human evaluation. The human face recognition experiment has also been carried out. The results show that this model can greatly improve the convergence of IGA, effectively alleviating human fatigue at the same time.