New Discrete Metaheuristic Approach for Large Scale Problem
International Journal of Machine Learning and Computing · 2022
Metaheuristic approaches have been widely used to solve large scale, complex global optimization problems.In this paper, a Hybrid simulated annealing based on discrete radius particle swarm optimization (H-DRPSOSA) with adaptive mutation is proposed.The proposed algorithm takes the advantage of the global search of the RPSO and the local search strategy of the SA algorithm to quickly generate good solutions.The paper also explains the framework design to solve the large scale multidimensional knapsack problems (LCO-MKPs).Additional, we present a random transfer mechanism for the feasible solution of the discrete search region.The proposed hybrid is compared to state-of-the-art solution techniques by applying them to the multidimensional knapsack dataset.Computational results demonstrate that the proposed algorithm is capable of producing competitive solutions.