ABSO: Advanced Bee Swarm Optimization Metaheuristic and Application to Weighted MAX-SAT Problem.

We introduce an advanced version of Bee Swarm Optimization metaheuristic (BSO) which is inspired from the foraging behavior of real bees. The objective of this work is to enhance the performances of BSO by subdividing the set of variables into groups covering disjointed sub-regions in the search space. To each sub-region is assigned a bee that performs a local search, and the search process is guided by the intensification and diversification principles. The subdivision of the set of variables is strongly dependent on the considered problem and aims at both reducing the execution time and maximizing the coverage of the search space. Our new approach called ABSO for Advanced Bees Swarm Optimization was applied to the weighted MAX-SAT and the comparison of experimental results showed that it outperforms the BSO algorithm.