NEW VARIATIONS OF THE GOLDEN BALL ALGORITHM WITH BIASED RANDOM KEYS: A COMPARATIVE INVESTIGATION APPLIED TO THE TRAVELING SALESMAN PROBLEM

This paper presents an investigation on nature-inspired and population-based metaheuristics. The main focus is on the introduction of three new variants of the Golden Ball (GB) algorithm. Variants include the integration of the Biased Random Key (BRK) technique into GB, the creation of parallel instances with BRKGB-I, and the simulation of football tournaments in BRKGB-WC and BRKGB-BFC. The experimental results, using instances of the Travelling Salesman Problem (TSP), show that the new variants offer a good relationship between solution quality and execution time, with emphasis on BRKGB-BFC, which obtained the best solutions in terms of quality.