Multi-directional bat algorithm for solving unconstrained optimization problems

In this paper, we propose a new hybrid algorithm for solving unconstrained global optimization problems by hybridizing the bat algorithm with multi-directional search algorithm (MDS). We call the proposed algorithm by multi-directional bat algorithm (MDBAT). In MDBAT algorithm, we try to overcome the slow convergence of the bat algorithm as a metaheuristic algorithm by invoking one of the promising direct search algorithm which is called MDS algorithm. The bat algorithm has a good ability to make exploration and exploitation search while the MDS has a good ability for accelerating convergence on the region of optimal response. In the beginning, the standard bat algorithm starts the search for number of iterations then the MDS algorithm starts its search from bat algorithm found so far. The combination between the standard bat algorithm and the MDS algorithm helps the MDS algorithm to start the search from a good solution instead of the random initial solution. The MDS algorithm can accelerate the search of the proposed algorithm instead of letting the algorithm running for more iterations without any improvement. We investigate the general performance of the MDBAT algorithm by applying it on 16 unconstrained global optimization problems and comparing it against 8 benchmark algorithms. The experimental results indicate that MDBAT is a promising algorithm and outperforms the other algorithms in most cases.