面向客户个性化需求的交互式遗传算法

Interactive genetic algorithm based on customer demand

针对交互式遗传算法(IGA)中的评价噪声问题,提出犹豫度的概念,建立犹豫度调整机制,并使用删除策略和修改策略来处理形成初始种群以及交叉、变异过程中产生的约束不满足个体.通过对汽车操控台的概念设计问题进行建模,建立人性化交互界面用以验证本论文提出的方法体系的先进性和合理性.实验表明,此求解算法能够有效的降低评价噪声,加速收敛,降低疲劳度,提高结果的满意度.

To solve the problem of evaluation noise in the interactive genetic algorithm （ IGA）, the concept of hesitancy degree is put forward, hesitancy degree adjustment mechanism is set up, and the deletion strategy and modification strategy are applied to handle the unsatisfied individuals generated in the process of initial population generation, crossover and mutation. By emulating the concept design of car console, the interactive interface with humanization is established to validate the advance and rationality of the method system brought forward in the paper. The experiment shows that the IGA can effectively reduce the evaluation noise, speed up the convergence, lower the fatigue degree and increase users＇ satisfaction about the result.