食物链算法及其与实编码遗传算法性能比较研究  

Research on Food-chain Algorithm and Compare Its Performance with Real-coded GA

作　　者：喻海飞[1] 汪定伟[2] 

机构地区：[1]东北大学工商管理学院,沈阳110004 [2]东北大学信息科学与工程学院,沈阳110004

出　　处：《系统工程理论与实践》2007年第6期143-148,共6页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(70431003;75103012;70571077)

摘　　要：根据人工生命突现集群和与环境进行动态作用的特点,同时基于生命系统中广泛存在的食物链现象,通过定义各级人工生命的局部活动规则,提出一种具有食物链形式的人工生命算法,并称之为食物链算法.详细阐述了食物链算法的定义、思想和算法流程;讨论了该算法中的生命代谢能量规则和人工生命活动邻域的变化规则;最后比较分析了食物链算法与基于实数编码的遗传算法在函数优化方面的性能.根据测试结果,食物链算法在函数优化方面具有明显的优势.另一方面,食物链算法具有很强的类生命特性,非常适合于类生物系统问题的优化求解.The characteristics of artificial life are emergence and dynamic interaction with the environment, and food-chain phenomenon appears to be an important and widespread phenomenon in the life system. Based on those ideas, we define the local action rules of artificial-life and propose a new type of artificial-life algorithm. It is referred as Food-Chaln Algorithm. Food-chain algorithm is a kind of artificial-life algorithms that achieve optimization computation by imitating the evolution of natural ecosystem and the information processing mechanism of natural organisms. In the next section, we firstly describe the definition, idea, and flow of food-chain algorithm. In the second, we discuss the energy rules of artificial-life metabolism and the transformation rules of movement range of artificial-life individuals. In the end, compared with real-coded genetic algorithm by standard functions, we find that food-chain algorithm is more predominant in function optimization. On the other hand, food-chain algorithm has lifelike traits, and it is fit to deal with the optimization problems of lifelike systems.

关 键 词：人工生命 食物链算法 突现集群 遗传算法 实数编码 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] F270.7[自动化与计算机技术—控制科学与工程]