交互式偏好满意优化在渔船技术经济论证中的应用

Evaluation of optimum fishing FRP ship types by an interactive preferred satisfactory optimization

针对在玻璃钢渔船技术经济论证中多目标优化的权重设计问题,提出一种交互式偏好权重的遗传算法,使决策者能够在多衡准满意优化中的当前群体中,将对于个体的偏好转换为对于权重数值的偏好关系,利用它们重新构造适应度函数,并作为在Pareto遗传算法的偏好优化方向。通过多次的进化后,决策者能够获取最满意的优化解。该方法避免了传统多目标船型论证中需要事先确定优化目标的权重且才能进行综合评判的缺陷,经玻璃钢渔船经济技术论证的计算实例论证,该方法是可行且有效的。

In view of the weight coefficients problem in multiobjective evaluation of optimum FRP fishing ship type, an interactive satisfactory method was proposed for Genetic Algorithm （GA） , which can make DM turn the preference for individual into for weight value, and used to reformulate evaluation function as the preferred optimization direction for Pareto GA, and finally DM can acquire the corresponding optimal satisfactory solution. The method avoids presetting the weight in traditional muhiobjective ship type evaluation, and an economic and technological evaluation of FRP ship type is used to demonstrate that the proposed method is of availability and efficiency.