一种引入偏好信息的多目标优化遗传算法  

A Multi-objective Optimization Genetic Algorithm Incorporating Preference Information

作　　者：申晓宁[1] 郭毓[1] 陈庆伟[1] 胡维礼[1] 

机构地区：[1]南京理工大学自动化学院,江苏南京210094

出　　处：《信息与控制》2007年第6期746-753,共8页Information and Control

基　　金：国家自然科学基金资助项目(60474034;60174019)

摘　　要：提出一种以交互方式引入决策者偏好信息的多目标优化遗传算法.该算法使用一种基于偏好信息的排序方法比较个体之间的优劣,并通过图形用户界面实现决策者与算法的交互.对算法的计算复杂度进行了理论分析,并进行了仿真实验.结果表明,所提算法具有较高的搜索效率,能够有效地求得期望区域内的折中解;特别当决策者的偏好发生变化时,算法能够快速做出响应,改变搜索范围,提供相应区域内的解.A multi-objective optimization genetic algorithm is proposed, which incorporates preference information of the decision maker in an interactive way. The algorithm makes use of a new ranking method based on the preference information to compare the individuals, and uses a graphical user interface to interact with the decision maker. Computational complexity of the algorithm is analyzed theoretically and simulation results indicate that the proposed algorithm can improve the searching efficiency, and can effectively find the trade-off solutions in the preferred region. Particularly, when the preference is changed, the algorithm can respond quickly to change the searching range and to find solutions in the corresponding region.

关 键 词：偏好 多目标优化 遗传算法 交互 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]