Random Search Method with Intensification and Diversification : Discrete Version
Kotaro Hirasawa, Hiroyuki Miyazaki, Jinglu Hu · QIR (Kyushu University Institutional Repository) (Kyushu University) · 2000
A new local search method RasID-D is proposed in this paper in order to improve the searching performance of discrete optimization problems. RasID-D is a discrete version of RasID which was developed for optimizing continuous optimization problems. RasID-D can achieve the high searching ability for the near optimal solutions without falling into local minima by iterating the intensification and diversification search. From simulations on Traveling Salesman Problem, it has been cleared that RasID-D can give better searching performances than the best local search method now being used.