Search diversification in ACO algorithms and its application

The author presents an approach to developing diversified ant colony optimization (ACO) algorithms, one of the most widely used combinatorial optimization methods. The proposed diversification in ACO algorithms is based on considering multiple options for extending the current solution fragment by incorporating several vertices of the problem graph into the route instead of just one, as is typically done. The ability of ants to foresee multiple search steps ahead increases the likelihood of avoiding suboptimal solutions and finding more accurate ones. This approach is applied to create metaheuristic algorithms for solving various combinatorial optimization problems. The results of computational experiments on a series of applied combinatorial optimization problems across different classes demonstrate the successful modification of the known ACO algorithms.