Abstract
The efficiency of genetic algorithms (GAs) depends mainly on crossover operator. The choice of good crossover operators leads to effective GA. There are many existing crossover operators in the literature. In this paper, we propose a modified version of sequential constructive crossover (SCX), named greedy SCX (GSCX), for solving the benchmark travelling salesman problem. We then compare the efficiency of the proposed crossover operator with greedy crossover, SCX and bidirectional circular SCX operators for solving the TSP on some benchmark TSPLIB instances. The comparative study shows that our proposed GSCX operator is the best among these crossover operators for the problem.