Adaptive Parameter Tuning with Double Layer Genetic Algorithm for Solving TSP

To address the issue of genetic algorithm (GA) performance heavily relying on parameter selection, this paper introduces a double-layer loop adaptive parameter tuning mechanism. This mechanism integrates the concept of GA as a meta-optimizer with adaptive GA, forming a nested two-layer GA structure. The outer GA, acting as the meta-optimizer, evaluates and adjusts the hyperparameters of the inner GA based on individual fitness changes. The inner adaptive GA solves the Traveling Salesman Problem (TSP) through genetic operations. Experimental results demonstrate that this method significantly enhances the algorithm's performance and efficiency by dynamically adjusting key parameters. Compared to traditional optimization algorithms, the proposed adaptive parameter tuning mechanism exhibits superior efficiency and stability in solving TSP.