New genetic algorithms for exploring design parameters for MEMS
MetadataShow full item record
The design of complex MEMS systems can be time consuming when many variables and geometric parameters are required. Proper exploration of the design space, which is required for finding good solutions have been a major challenge. This thesis applies novel genetic algorithm methods (AVGM, RBAM and MORBAM) to the design of micro- electromechanical systems (MEMS). The main objectives of the algorithms, which are introduced in this thesis is better identification of high performance region for MEMS geometry design and faster computational time. The Average-Mixture (AVGM) and the Random Based Average Mixture (RBAM) Genetic Algorithm methods are applied to the single objective problems while the Multi-Objective Random Based Method is applied to the multi-objective problems. The main advantages of the methods over the traditional genetic algorithm methods are their ability to identify high performance regions while maintaining diversity by exploring the search space efficiently. These algorithms provide many good results, which are diverse in nature.