Fast evolutionary algorithms

  • Authors:

  • Xin Yao;Yong Liu;Ko-Hsin Liang;Guangming Lin

  • Affiliations:

  • School of Computer Science, The University of Birmingham, Edgbaston, Birmingham B15 2TT, UK;Evolvable Systems Laboratory, Computer Science Division, mbox 1501, Electrotechnical Laboratory, 1-1-4 Umezono, Tsukuba, Ibaraki 305-8568, Japan;School of Computer Science, University College, The University of New South Wales, Australian Defence Force Academy, Canberra, ACT, Australia 2600;School of Computer Science, University College, The University of New South Wales, Australian Defence Force Academy, Canberra, ACT, Australia 2600

  • Venue:
  • Advances in evolutionary computing
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

This chapter discusses a number of recent results in evolutionary optimization. In particular, we show that the search step size of a variation operator plays a vital role in its efficient search of a landscape. We have derived the optimal search step size of mutation operators in evolutionary optimization. Based on this theoretical analysis, we have developed several new evolutionary algorithms which outperform existing evolutionary algorithms significantly on many benchmark functions.Most of the existing work in evolutionary optimization concentrates on different variation (i.e., search) operators, such as crossover and mutation. However, there may be a better way to solve a complex problem by transforming it into a simpler one first and then solving it. The key issue here is how to approximate the problem without changing the nature of the problem (i.e., the optima we wish to find). This chapter will present the latest results on landscape approximation and hybrid evolutionary algorithms.