摘要：

为提高交互式遗传算法的优化效率,提出一种基于进化个体混杂型适应值的交互式遗传算法。设计适应值不确定度计算方法,分析适应值噪声特性。在此基础上,根据偏好不确定性与适应值噪声的内在联系,划分出单一数值与区间数值2种适应值类型,并分别建立相应数学模型,修正个体适应值,使其同时参与进化优化,从而生成符合用户心理需求的设计,达到高效优化目的。在便携式酒壶设计系统上的应用结果表明,与IGA-IIF和T-IGA算法相比,该算法不仅进化代数相对较少,而且每代可以获取更多的互异个体数目,具有较高的效率。

关键词: 遗传算法, 适应值, 交互式, 噪声, 进化个体

Abstract:

To improve the optimize efficiency of interactive genetic algorithm,this paper proposes an interactive genetic algorithm based on hybrid fitness of evolutionary individuals.The calculation method of fitness uncertainty is designed while analyzing the characteristics of fitness noise,uncertainty analysis is performed on individual adaptive values of the user evaluation,and two kinds of adaptive value types including the single value type and the interval value type are divided on the basis of the uncertainty degree of the adaptive value.It builds a corresponding mathematical model by aiming at the adaptive value types,the individual adaptive values are corrected,and participate in the subsequent evolution.The two kinds of corrected adaptive values simultaneously participate in the evolution optimization,the design conforming to the psychological need of the user is expected to be generated,and the goal of efficient optimization is achieved.The proposed algorithm is applied to a portable wine pot design system.Experimental results confirm that,compared with IGA-IIF algorithm and T-IGA algorithm,it has advantages in improving optimization efficiency and alleviating user fatigue while improving its efficiency in exploration and practical application.

Key words: Genetic Algorithm(GA), fitness, interactive, noise, evolutionary individual

中图分类号: 

• TP273