Multi-Objective Ant Colony Optimization Based on the Physarum-Inspired Mathematical Model for Bi-Objective Traveling...

Area of Science:

• Operations Research
• Computational Intelligence
• Metaheuristics

Background:

• The Bi-objective Traveling Salesman Problem (bTSP) is a complex combinatorial optimization problem with significant real-world applications.
• Existing Multi-objective Ant Colony Optimization (MOACO) algorithms often face challenges with premature convergence, limiting their effectiveness for bTSPs.

Purpose of the Study:

• To propose an enhanced MOACO strategy for solving bTSPs by optimizing pheromone matrix initialization.
• To leverage the Physarum-inspired Mathematical Model (PMM) to guide the initialization process and improve ant search capabilities.

Main Methods:

• Integration of a Physarum-inspired Mathematical Model (PMM) for initializing the pheromone matrix in MOACO algorithms.
• Development of an optimized algorithm, termed iPM-MOACOs, to enhance pheromone concentration on shorter paths.
• Conducting a series of experiments to evaluate the performance of the proposed iPM-MOACOs against standard MOACOs.

Main Results:

• The iPM-MOACOs demonstrated an improved ability to avoid premature convergence compared to traditional MOACOs.
• Experimental results indicated that the proposed strategy enhances the search efficiency of ants by reinforcing shorter paths.
• The optimized algorithms achieved better compromise solutions for bTSP instances.

Conclusions:

• The proposed pheromone initialization strategy using PMM effectively enhances MOACO performance for bTSPs.
• iPM-MOACOs offer a superior approach to finding high-quality compromise solutions for bTSP.
• This research contributes a novel optimization technique to the field of multi-objective optimization using bio-inspired algorithms.