Dynamic Constrained Multiobjective Evolutionary Algorithms And Their Application In Antenna Design

Constrained optimization problems(COPs)are commonly encountered in the disciplines of science and engineering application.During the past decades,the researchers have widely adopted evolutionary algorithms(EAs),which are effective stochastic search techniques inspired by nature,to solve COPs.Multiobjective optimization has been one of the hottest research topics in the area of evolutionary computation since 1980's,and many mature methods have been proposed.The use of multiobjective optimization techniques based on Pareto dominance to solve COPs is an exciting avenue for research.The antenna design is a kind of optimization problems which have nonlinear functions,many variables,complex constraints,it needs the high-performance constrained optimization evolutionary algorithms to be solved.Considering the above,the project of “Dynamic constrained multiobjective evolutionary algorithms and their applications in antenna design” is proposed in this paper.The main works are as follows:1.A dynamic constrained many-objective optimization evolutionary algorithm(DCMa OEA)is presented to solve constrained optimization problems.The COP is converted an equivalent constrained many-objective optimization problem.So many-objective optimization techniques can be incorporated into the method to maintain diversity.Besides,the dynamic handling mechanism is adopted to deal with constraints.In the beginning of the algorithm,the constraints boundaries are broadened so as to make the whole population feasible,then they shrink gradually to the original boundaries with the evolution proceeding.This way the manyobjective optimization algorithm is able to work well without the affection of the constraints.DCMa OEA uses the differential evolution as a search engine and the reference-point-based nondominated sorting approach as the selection of individuals to generate the next population.The proposed algorithm is tested on 60 benchmark functions,and compared with some state-of-the-art algorithms.From the results,DCMa OEA has shown its good performance to deal with COPs.2.Based on the dynamic constrained handling mechanism and the multiobjective optimization technology,three improved algorithms are proposed,they are hybrid constraint handling mechanism evolutionary algorithm(HCEA),DCMa OEA with computational resource allocation(DCMa OEA-CRA)and DCMa OEA with parameters leaning(DCMa OEA-PL).These algorithms are tested on CEC 2006 benchmark functions,the experimental results demonstrate that they perform better than the original algorithm.3.The proposed algorithm is applied in designing evolutionary antenna.DCMa OEA is employed to optimize two antennas,a low-profile wide-beamwidth circularlypolarized antenna and an S-band medium gain antenna.The effectiveness of DCMa OEA is verified through simulated and measured results.The innovation point in this paper is the use of many-objective optimization techniques for solving COPs.Among other studies,problems generally are considered as biobjective optimization problems: an original objective and a violation objective formulated from constraints functions.These constraints usually have different characteristics,if it only use the sum of them,some information will be lost.DCMa OEA adopts many-objective optimization to deal with constraints,it is more suitable to solve complex COPs.