基于斜率淘汰策略的多目标演化算法  

Slope-based Elimination Multi-Objective Evolutionary Algorithm

作　　者：郑彬彬[1] 李元香[1] 

机构地区：[1]武汉大学计算机学院,湖北武汉430072

出　　处：《计算机应用研究》2006年第6期64-66,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(60473014);国家博士学科点科研基金资助项目(20030486049)

摘　　要：提出一种新的多目标演化算法———基于斜率淘汰策略的多目标演化算法。利用基于斜率的淘汰策略,在演化过程中能以较低的时间复杂度更新精英空间、保存精英个体(E litist),且取得的解数量大,既保证了近似解集对Pareto前沿的逼近,又很好地保持了解集分布的均匀性。对于一些代表性的BenckMark问题,数值试验都取得了非常好的效果。This paper puts forward a new multi-objective evolutionary algorithm： slope-based elimination multi-objective optimization algorithm, which can refresh the population in a lower time complexity in the process of evolvement, and save the elitist set. And the number of elitist set is large. So that the population could approach the Pareto front approximately. At the same time, this new algorithm keeps the population in a equal distribution. The experiments show that the algorithm has very good performance.

关 键 词：多目标优化 多目标演化算法 PARETO最优集 斜率淘汰策略 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]