Effective Hybridization of Biogeography Based Optimization and Simulated Annealing

Abstract

Since the optimization process constitutes a great step in solving complex real world problems, the development of novel optimization algorithms is one of the growing interest topics that attracted researchers in the recent decades. This paper presents the hybridization of bio-inspired Biogeography Based Optimization (BBO) algorithm and physics-inspired Simulated Annealing (SA) algorithm, into a new variant called BBO-SA. The proposed algorithm uses the concepts of the SA to enhance the diversity of BBO solutions which in turn improves the obtained solution. For validating the performance of BBO-SA, it is compared to that of BBO algorithm in solving a set of thirteen complex benchmark functions. Validation results prove the superior performance of the proposed BBO-SA algorithm over the BBO algorithm in solving complex function in terms of escaping from local optima and reaching near optimal solution in lower execution times. Besides, the proposed algorithm is applied to solve a very challenging problem denoted as the RFID Reader Deployment Problem (RRDP). Such problem can be solved by finding the optimal distribution of the RFID readers which fulfils the set of RFID planning objectives. A comparison is held between the BBO-SA algorithm and other optimization algorithms on a large RFID model. Simulation results verified the superiority of the algorithm over the compared ones for solving the RRDP with satisfying the deployment objectives.