A Novel Approach for Reducing the Computational Cost of GA in Synthesizing Low Sidelobe Unequally Spaced Linear Arrays

摘要

This article presents two approaches to reduce the computational cost of genetic algorithm (GA) in generating low sidelobe linear arrays by optimizing the element positions subject to the multiple design constraints of the number of elements, the array aperture dimension, and the minimum interelement spacing. Many experimental results indicate that only considering a few edge elements on both the ends of the array will have the same effect on the low sidelobe pattern synthesis with the case of taking all the elements into account. Firstly, based on this concept, the equation for calculating the number of the edge elements is found by using the statistical method. With this equation the number of gene variables of the chromosome of GA in the evolutionary process is considerably reduced. Undoubtedly, the computational cost of GA is effectively reduced. Then, at the same time of presenting this equation, a new manner of individual description is introduced. Using it and modifying the optimization process of GA, the size of the searching space of GA is reduced and the infeasible solutions during the evolutionary process are avoided. Thus the computational cost of GA is further reduced. Example arrays with low sidelobe pattern are synthesized to assess the efficiency and robustness of the proposed GA. The simulated results demonstrate that the computational cost of GA is considerably reduced and preferable sidelobe levels are obtained compared with that of previous works. antenna arrays; nonuniformly spaced arrays; side lobe suppression; genetic algorithms (GA)