A Multi-parent Hybrid Order and Cost-based Sequential Constructive Crossover of Genetic Algorithm for the Traveling Salesman Problem
Zhida Liu, Biao Di, Juan K. Lin · 2023
We investigate Genetic Algorithm (GA) where multi-parent crossovers are involved to replace the traditional two-parent crossover operation in permutation-based problems. And we propose a new multiple-parent hybrid order and cost-based sequential constructive crossover (MPHOSCX) for the Traveling Salesman Problem (TSP). More genetic high-quality information is passed on through multiple parents, cost-based crossover ensures the efficiency of the operation. Further, a nearest neighbor inverse operation is used to enhance the overall capability of exploitation. Experimental results confirm the merits of using the multi-parent hybrid crossover. And the comparison results with other multi-parent GA and GAs for TSP demonstrate the superiority of the proposed operators.