Searching for backbones - An efficient parallel algorithm for the traveling salesman problem

被引:45
作者
Schneider, J [1 ]
Froschhammer, C [1 ]
Morgenstern, I [1 ]
Husslein, T [1 ]
Singer, JM [1 ]
机构
[1] UNIV ZURICH,INST PHYS,CH-8057 ZURICH,SWITZERLAND
关键词
optimization; parallel; Monte Carlo; threshold accepting; TSP; backbone; degeneracy;
D O I
10.1016/0010-4655(96)00062-8
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
The Traveling Salesman Problem (TSP) plays an important role in Operations Research, Applied Mathematics and Computational Physics. We investigated it using a stochastic approach. Studying several solutions of a special TSP we found that many parts of a good solution are the same in all other good solutions for this problem. In this paper we discuss an efficient parallel method to reduce the TSP to a smaller one by finding these backbones and eliminating them to get even better solutions in a very short time and a few observables of interest corresponding to this parallel approach.
引用
收藏
页码:173 / 188
页数:16
相关论文
共 24 条
[21]   DETERMINATION OF CANDIDATE STRUCTURES FOR SIMPLE IONIC COMPOUNDS THROUGH CELL OPTIMIZATION [J].
SCHON, JC ;
JANSEN, M .
COMPUTATIONAL MATERIALS SCIENCE, 1995, 4 (01) :43-58
[22]  
STADLER PF, 1991, LANDSCAPE TRAVELING
[23]   SCALING FEATURES IN COMPLEX OPTIMIZATION PROBLEMS [J].
TAFELMAYER, R ;
HOFFMANN, KH .
COMPUTER PHYSICS COMMUNICATIONS, 1995, 86 (1-2) :81-90
[24]  
TAFELMAYER R, 1994, ADAPTIVE SCHEDULES E