大学サイト   学部サイト   学科サイト   EnglishPages
 
旧尾家・塚本研究室です.尾家祐二 現・九州工業大学学長の情報は こちら.
   
研究業績


Non-swarm Intelligence Search Algorithm Based on the Foraging Behaviors of Fruit Flies
著者
  Kei Ohnishi, Akihiro Fujiwara, and Koeppen Mario
 
雑誌名/会議名
  2016 IEEE Congress on Evolutionary Computation (CEC2016), pp.1509-1516, Vancouver, Canada, July 24-29, 2016. 2016 年 7 月
 
アブストラクト
  We focus on a fruit fly (Drosophila) that is one of non-social insects independently autonomously foraging unlike social insects and investigate the foraging behaviors of fruit flies, and then propose a new population-based and non-swarm intelligence search algorithm mimicking the behaviors. The fact we learn is that a histogram of duration for which a fruit fly stays in a food source follows a power-law. A power-law distribution holds a scale-free property, so that the behaviors of fruit flies observed in the fact are called ``scale-free behaviors’’, which the proposed algorithm mimics. We examine the effect of the scale-free behaviors on search performance by applying the proposed algorithm to 28 test functions with two real parameters. The examination results show that the scale-free property related to a balance between exploration and exploitation indeed affects the search performance and particular parameter values for fixing the scale-free property yield better search performance for most test functions used. The contribution of the paper is not to provide a new competent search algorithm for static optimization problem but to introduce the novel temporal property of individual behaviors, the scale-free property, which would be expected to be a new aspect to understand search algorithms.
 
キーワード
 
 
記述言語
  English
 
 

web_admin [ at ] infonet.cse.kyutech.ac.jp /  Copyright (c) 2016 Oie & Tsukamoto Laboratory, All Right Reserved.