Modified gorilla troops optimization for the quadratic assignment problem

Abstract

Balancing exploration and exploitation remain a fundamental challenge in artificial intelligence-based optimization, particularly when addressing discrete combinatorial problems such as the quadratic assignment problem (QAP). The gorilla troops optimizer (GTO), inspired by the collective social behavior of gorillas, has shown promising results in continuous domains but faces limitations when directly applied to discrete optimization. To address this, the present study introduces a modified gorilla troops optimizer (MGTO), a novel discrete adaptation designed specifically for the QAP. The proposed MGTO strategically integrates a swapping-based diversification mechanism to enhance exploration within discrete solution spaces, while a modified uniform crossover operator promotes effective exploitation of high-quality solutions. Extensive experiments on benchmark instances from the quadratic assignment problem library (QAPLIB) show that MGTO achieves superior convergence behavior and solution quality compared with several state-of-the-art algorithms. These results demonstrate MGTO’s capacity to maintain a balanced equilibrium between exploration and exploitation, effectively navigating complex discrete landscapes to yield high-quality solutions with strong computational efficiency.