A Novel Hybrid Optimization Algorithm Based On BBO And PSO And Its Application

In the past few decades,intelligent optimization algorithms based on evolutionary computation(EA)have attracted more and more attention,which is due to the fact that this kind of algorithms can solve a mass of problems that can't be solved by traditional optimization methods.With many research and attempts by researchers,now the EA optimization algorithms have been successfully employed in various areas of science and engineering and brought prominent benefit.A large number of heuristic EAs have been emerging for the sake of their superiority,in addition,the emergence of a classical EA can generally result in many modified algorithms based on it in order to improve its search capacity and solve specific problems.There are many methods which can be used to modify a EA,such as changing its parameters and adopting some novel search strategies,etc.However,a single EA has finite search capacity after all,as a result,it's very necessary to combine different EAs with a well hybridization strategy and compose hybrid EAs,which has become a hotspot.In general,hybrid EAs can synthesize the advantages from the constituent EAs and have distinct superiority in search and optimization compared with single EAs.This research focus on a novel heuristic EA named BBO and a classical swarm intelligence EA called PSO,a hybrid EA named BBO-PSO will be proposed by combining the above two EAs after analyzing their search mechanism and information flow.This hybrid EA can not only coordinate local search with global search but also balance exploitation with exploration through synthesizing the advantages of BBO and PSO.The tests on benchmark functions show that BBO-PSO has powerful search ability with satisfactory robustness and high efficiency.In order to prove its availability on engineering,BBO-PSO will be used to design 2-D IIR digital filters by transforming the design problem into a single-objective optimization problem and building its mathematical model.Simulation results show that BBO-PSO has better performance compared with existing design methods on this problem,which further reveals the potential applications of this hybrid EA.In the end of this paper,a modified algorithm called IBBO-PSO is proposed to further improve the computational efficiency and optimization capability of BBO-PSO.