Optimal tree for Genetic Algorithms in the Traveling Salesman Problem (TSP)

TL;DR

This paper introduces an optimal tree method to generate initial populations in genetic algorithms for the TSP, aiming to improve efficiency by excluding poor parent tours and analyzing the gauge's characteristics.

Contribution

It proposes using an optimal tree as a gauge to generate initial populations in GAs for TSP, enhancing speed and benchmarking capabilities.

Findings

Optimal tree effectively filters bad parent tours.

Speed-up in genetic algorithm convergence.

Analysis of gauge characteristics and complexity.

Abstract

In this paper, the author proposes optimal tree as a "gauge" for the generation of the initial population at random in the Genetic Algorithms (GA) to benchmark against the good and the bad parent tours. Thus, without having the so-called bad parent tours in the initiate population, it will speed up the GA. The characteristics of the gauge (algorithm, complexity time, trade-off, etc.) will be discussed in this paper as well.