A Fuzzy Logic for Parameter Adaptation in Ant Colony Optimization Approach

Abstract

The purpose of this study is to produce an "adaptive Ant Colony System" in order to establish a balance between exploitation and exploration in terms of solving the Travel salesman problem. First we do a detailed investigation of several Ant Colony System Algorithm�s parameters. Second we will incorporate into the algorithm a Fuzzy Logic Controller, which will be utilized to alter the settings based on the algorithm�s reliable performance metrics. The parameter customization will proceed throughout the execution of the algorithm, providing for a dynamic parameter settings depending on the algorithm�s current performance. The adaptive algorithm will be examined on a set of TSP problems of varying sizes, and the results will be compared to those obtained using the standard algorithm and other studies in the same manner.