Antenna Optimization Based On Domain Decomposition Method And Particle Swarm Optimization

CEM have beenrapidly developedand widely used in engineering and academic nowadays.However, theelectromagnetic analyzeefficiency of complicated large scale and multi-scale objectis still a matter of concern.The application based on integral equation domain decomposition method(DDM) and particle swarm optimization(PSO) is studied systematically to deal with antenna optimization inthis dissertation.The proposednon-conformaldomain decomposition methodprovidesa computationally efficient and effective way to analyze the electromagnetic problemsofcomplicatedobject.According tothe strategy of divide-and-conquertheorem, DDM first decomposes the original problem intoseveral independentsubdomains. Subsequently, each of the sub-regions can be solved separately and then we iterate between them to get the final numerical solution.Moreover, cooperated with PSO, DDM is applied to the design of stealthstructures and antennadistribution.First, the basic optimization methods,typicallythe intelligent optimization algorithms, are presented. Wethen discuss the principle, structure and improvement of the PSO as the foundation of further optimization design.To describe the theoryand realization of domain decomposition method, the formulation of DDMis derived based on integral equation.After divided into different closed subdomains and meshed flexibly, the whole object can be solved separately. The adjacentsubdomains are connectedby the transmission condition. Especially thanks to the non-conformaltechnique and projection judgment,DDMprovides unprecedentedflexibility and convenience.This renders geometricallycomplicated and alterable subdomain problems tractable.At last, an application by combiningDDM and PSO isproposed tooptimize the stealth and distribution of carrierantenna.This method makes it possible toavoid re-mesh during the design process, and provides us with a new efficient and accurate tool for theoptimization of complicated multi-scale electromagneticproblems.