Big-valley Explorer: A Framework of Real-coded Genetic Algorithms for Multi-funnel Function Optimization

• Uemura Kento

  Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology; Research Fellowship for Young Scientists (DC1) of the Japan Society for the Promotion of Science

• Kinoshita Shun-ichi

  Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology (Present: Cyberscience Center, Tohoku University)

• Nagata Yuichi

  Education Academy of Computational Life Sciences, Tokyo Institute of Technology

• Kobayashi Shigenobu

  Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology (Present: Professor Emeritus, Tokyo Institute of Technology)

• Ono Isao

  Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology

Bibliographic Information

Other Title

• 大域的多峰性関数最適化のための実数値GAの枠組みBig-valley Explorerの提案

Abstract

This paper proposes a new framework of real-coded genetic algorithms (RCGAs) for the multi-funnel function optimization. The RCGA is one of the most powerful function optimization methods. Most conventional RCGAs work effectively on the single-funnel function that consists of a single big-valley. However, it is reported that they show poor performance or, sometimes, fail to find the optimum on the multi-funnel function that consists of multiple big-valleys. In order to remedy this deterioration, Innately Split Model (ISM) has been proposed as a framework of RCGAs. ISM initializes an RCGA in a small region and repeats a search with the RCGA as changing the position of the region randomly. ISM outperforms conventional RCGAs on the multi-funnel functions. However, ISM has two problems in terms of the search efficiency and the difficulty of setting parameters. Our proposed method, Big-valley Explorer (BE), is a framework of RCGAs like ISM and it has two novel mechanisms to overcome these problems, the big-valley estimation mechanism and the adaptive initialization mechanism. Once the RCGA finishes a search, the big-valley estimation mechanism estimates a big-valley that the RCGA already explored and removes the region from the search space to prevent the RCGA from searching the same big-valley many times. After that, the adaptive initialization mechanism initializes the RCGA in a wide unexplored region adaptively to find unexplored big-valleys. We evaluate BE through some numerical experiments with both single-funnel and multi-funnel benchmark functions.

Journal

• Transaction of the Japanese Society for Evolutionary Computation

  Transaction of the Japanese Society for Evolutionary Computation 4 (1), 1-12, 2013

  The Japanese Society for Evolutionary Computation

Keywords

• function optimization
• multi-funnel function
• real-coded genetic algorithms
• ISM

Details

• CRID: 1390001205366429312
• NII Article ID: 130004965145
• DOI: 10.11394/tjpnsec.4.1
• ISSN: 21857385
• Text Lang: ja
• Data Source:
  • JaLC
  • CiNii Articles
  • KAKEN
• Abstract License Flag: Disallowed