CROSSOVER ON INTENSIVE SEARCH AND TRAVELING SALESMAN PROBLEM

This paper describes a new crossover opera tor for the traveling salesman problem, the greedy selection crossover operator (GSX for short), which is designed for path representation and performed. at gene level. It can utilize both of local precedence and global precedence relationship among genes to perform intensive search among solution space to reproduce an improved offspring. It has been compared with other well known crossover operators and the results show that GSX operator is clearly superior to traditional operators. The proposed algorithm can be readily applied to other combinatorial optimization problems if a measure is properly defined based on the global information for a given problem.