Research Topic (nur in Englisch)

Abstract

Genetic algorithms (GAs) are well-known heuristic algorithms and have been applied to solve a variety of complicated problems. When adopting GA approaches, two important issues — selection pressure and population diversity — must be considered. We present a novel mating strategy, called tabu genetic algorithm (TGA), which harmonizes these two issues by integrating tabu search (TS) into GA’s selection.

TGA incorporates the tabu list to prevent inbreeding so that population diversity can be maintained, and further utilizes the aspiration criterion to supply moderate selection pressure. An accompanied self-adaptive mutation method is also proposed to overcome the difficulty of determining mutation rate, which is sensitive to computing performance. Experimental results on TSP indicate that TGA can achieve harmony between population diversity and selection pressure. Comparisons with GA, TS, and hybrids of GA and TS further confirm the superiority of TGA in terms of both solution quality and convergence speed.

Selected Publications

• S. C. Lin and C. K. Ting. A new approach for detection of dimensions set in mechanical drawing.
  Pattern Recognition Letters, Elsevier Science,  no.18, pp. 367-373, 1997.
• C. K. Ting, S. T. Li and C. N. Lee. On the harmonious mating strategy through tabu search.
  Information Sciences, Elsevier Science, (to appear)
• C. K. Ting, C. N. Lee and S. T. Li. A novel hybrid optimization algorithm based on genetic algorithm and tabu search.
  Proceedings of International Computer Symposium, pp. 157-162, 2000.
• C. K. Ting, S. T. Li, and C. N. Lee. TGA: a new integrated approach to evolutionary algorithms.
  IEEE Congress on Evolutionary Computation (CEC2001), Seoul, Korea, pp. 917-924, 2001.
• S. T. Li, C. K. Ting, and C. N. Lee. Maintenance scheduling of oil storage tanks using tabu-based genetic algorithm.
  4th IEEE International Conference on Tools with Artificial Intelligence (ICTAI'02), Washington D.C., USA, pp. 209-215, 2002.
• C. K. Ting. Design and analysis of multi-parent genetic algorithms.
  PhD-Thesis, University of Paderborn, 2005.