A hybrid discrete particle swarm algorithm for hard binary CSPs

  • Authors:

  • Qingyun Yang;Jigui Sun;Juyang Zhang;Chunjie Wang

  • Affiliations:

  • College of Computer Science and Technology, Jilin University, Changchun, China;College of Computer Science and Technology, Jilin University, Changchun, China;College of Computer Science and Technology, Jilin University, Changchun, China;Basic Sciences of ChangChun University of Technology, ChangChun University of Technology, Changchun, China

  • Venue:
  • ICNC'06 Proceedings of the Second international conference on Advances in Natural Computation - Volume Part II
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

The discrete particle swarm algorithm for binary constraint satisfaction problems (CSPs) is analyzed in this paper. The analysis denotes that ϕ1 and ϕ2 are set to 0 may be a heuristic similar to min-conflict heuristic. The further observation is the impact of local best positions. A control parameter pb is introduced to reduce the effect of the local best positions. To improve the performance, simulated annealing algorithm is combined with the discrete particle swarm algorithm, and the neighborhood exploring in simulated annealing is carried out by ERA model. Eliminating repeated particles and Tabu list avoiding cycling are also introduced in this paper. Our hybrid algorithm is tested with random constraint satisfaction problem instances based on phase transition theory. The experimental results indicate that our hybrid discrete particle swarm algorithm is able to solve hard binary CSPs.