Using Cooperation Approaches at Different Levels of Artificial Bee Colony Method

In this work, a Multi-Level Artificial Bee Colony (called MLABC) for optimizing numerical test functions is presented. In MLABC, two species are used. The first species employs n colonies where each of them optimizes the complete solution vector. The cooperation between these colonies is carried out by exchanging information through a leader co lony, which contain s a set of elite bees. The second species uses a cooperative approach in which the complete solution vector is divided to k sub-vectors, and each of these sub-vectors is optimized by a colony. The cooperation between these colonies is carried out by compiling sub-vectors into the complete solution vector. Finally, the cooperation between two species is obtained by exchanging information. The propo sed algorithm is tested on a set of well-known test functions. The results show that MLABC algorithm provides eff iciency and robustness to solve numerical functions. Keywords—Artificial bee co lony, cooperative artificial bee colony, multilevel cooperation.