Metaheuristic Algorithms and Polynomial Turing Reductions: A Case Study Based on Ant Colony Optimization

Nowadays, there is an increasing dependence on metaheuristic algorithms for solving combinatorial optimization problems. This paper discusses various metaheuristic algorithms, their similarities and differences and how Ant Colony Optimization algorithm is found to be much more suitable for providing a generic implementation. We start with the solution for Travelling Salesman Problem using Ant Colony Optimization (ACO) and show how Polynomial Turing Reduction helps us solve Job Shop Scheduling and Knapsack Problems without making considerable changes in the implementation. The probabilistic nature of metaheuristic algorithms, especially ACO helps us to a greater extent in avoiding parameter fine-tuning. Through Sensitivity analysis we find that ACO exhibits better resilience to changes in parameter values in comparison to other metaheuristic algorithms.