A New Metaheuristic Approach Combining ACO and H-method
Abs. of Int. Conf "Discrete and Global Optimization" (July 31–August 2, 2008, Yalta, Ukraine). – Kyiv, 2008. – P. 18.
Combinatorial optimization problems and approximate algorithms of their solving are considered in the paper. An approach to COP formalization is suggested, which allows not only to distinguish single COP classes, but also to classify optimization problems in general. A new metaheuristic method of solving NP-hard combinatorial optimization problems is proposed, which is based on two population algorithms – ant colony optimization (ACO) and H-method. ACO is bio-inspired example of swarm intelligence and has been successfully applied to a number of hard combinatorial optimization problems. H-method is a population-based metaheuristic, applying during the search process specially de?neds egments. The method shares remote similarities with a well-known nondifferential continuous optimization method – the Nelder-Mead method. The presented method was tested on benchmark instances of multidimensional assignment problem.