The Application Of Improved Particle Swarm Optimization Algorithm In The MEMS Component Design

Particle Swarm Optimization algorithm (PSO) as an intelligenceoptimization algorithm, which derived from bird foraging behavior, is anoptimization tool based on iterations. Because of its simple structure andconvenience, and other advantages, PSO has been widely applied in all kinds offields, such as industrial control, artificial neural networks, circuit design, andso on.Micro-Electro-Mechanical Systems (MEMS) component is a new type ofcomponents with micro-electro-mechanical systems technology. Along with therapid development of wireless technology, especially in the radar and radiofields, the device with miniaturization, low-power, and integration has receivedmuch attention. Considering of MEMS component advantages and itsapplication in wireless technology，some research has been widely studied.There are two main methods in the traditional design of MEMS research:formula and software simulation. Considering of the disadvantages such as largecomputation and time-consuming and so on in the traditional design, a new design method, improved Particle Swarm optimization, is been widely appliedin the MEMS devices design.In this paper, Hybrid Adaptive Mutation Particle Swarm Optimization(HAMPSO) improves a lot on the basis of basic PSO and is successfully appliedto design MEMS devices. The main contents are as follows:(1) Hybrid Adaptive Mutation Particle Swarm Optimization (HAMPSO).Basic PSO has many advantages such as premature phenomenon, poor precision,loss of particle diversity and so on. Considering the disadvantages of basic PSO,propose HAMPSO by introducing artificial fish swarm algorithm (AFSA) andcauchy mutation algorithm. It is tested by typical functions, which shows highercalculating precision, faster speed and the ability of avoid being premature.(2) The optimization design of MEMS devices. HAMPSO combined withADS and HFSS is used to optimize structure parameters of MEMS devices. Theresults have showed that the new method can save design time and speed up thedesign process without losing the performance of MEMS devices.