A Constrained GA for Electro-Thermal Microactuator Design

• Main Authors: Po-Hung Lin, 林柏宏
• Other Authors: Chien-Jong Shih
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/46002831721249149236

碩士 === 淡江大學 === 機械與機電工程學系碩士班 === 99 === A conventional genetic algorithm (GA) is called simple GA that is a nature optimization algorithm without constraint handling. Several methods belong to attached strategy where constraints are transformed to a part of pseudo objective function to execute constraint-handling. From the inspiration of double helix chromosome and its behavior, a constrained GA naturally should contains the ability of handling constraints during the evolutionary process, that is presented in the thesis. Three operators: recombination, transplantation and mutation, are provided to construct such an evolutionary algorithm to deal with constrained GA. Numerical examples show that the proposed constrained GA can be satisfied by the verification of precision, efficiency and robustness. Electro-thermal microactuator has been explored in this work for engineering analysis and design. A satisfying finite element model has been constructed, and then can be utilized for optimum design. Single-arm and double-arm Electro-thermal microactuator are analyzed and discussed and provided as the design applications for constrained GA optimization. The feature of the proposed design model includes an extra design variable - voltage, and design constraint of temperature limitation. To do the suitable temperature control, the optimum input voltage can be obtained in the proposed design method.