A new approach in dynamic traveling salesman problem: a hybrid of ant colony optimization and descending gradient

TL;DR

This paper introduces a hybrid algorithm combining ant colony optimization and gradient descent to improve dynamic traveling salesman problem solutions, preventing premature convergence and achieving better route optimization.

Contribution

It presents a novel hybrid approach that integrates ACO and gradient descent, enhancing solution quality for the DTSP compared to previous methods.

Findings

Significantly improved route optimization results.

Prevents premature convergence and local optima trapping.

Outperforms some existing methods in DTSP solutions.

Abstract

Nowadays swarm intelligence-based algorithms are being used widely to optimize the dynamic traveling salesman problem (DTSP). In this paper, we have used mixed method of Ant Colony Optimization (AOC)and gradient descent to optimize DTSP which differs with ACO algorithm in evaporation rate and innovative data. This approach prevents premature convergence and scape from local optimum spots and also makes it possible to find better solutions for algorithm. In this paper, we are going to offer gradient descent and ACO algorithm which in comparison to some former methods it shows that algorithm has significantly improved routes optimization.