Parallel island model based hybrid black widow algorithm for solving the traveling salesman problem

The Traveling Salesman Problem (TSP), one of the fundamental problems in the field of combinatorial optimization, remains a focal point for researchers due to its NP-Hard structure and complex solution space. In this study, a Parallel Island Model Based Hybrid Black Widow Optimization (PH-BWO) algorithm, developed based on the Black Widow Optimization (BWO) algorithm, is proposed for the effective solution of the TSP. The proposed method aims to overcome fundamental limitations by being supported by genetic operators, a scale-sensitive parameter system, and dynamic local search strategies. Through the island architecture designed in the study, exploration is ensured, while the cannibalism mechanism has been refunctionalized as a directed selection pressure, thereby enhancing the algorithm's exploitation capability.The performance of the algorithm was tested on TSPLIB datasets of various scales. The findings indicate that PH-BWO consistently reaches the best-known values, particularly in small and medium-scale problems. In large-scale problems, it exhibited a stable structure and approached the best-known values with low deviation rates. Analyses confirmed that the method is superior in terms of speed and solution quality to the Hybrid-BWO and SDABWO models found in the literature. This study presents an original approach addressing the BWO hybrid for TSP within a holistic framework in the Turkish literature and establishes a theoretical foundation for future studies.