Bi-directional evolutionary level set method for topology optimization

Abstract

A bi-directional evolutionary level set method for solving topology optimization problems is presented in this article. The proposed method has three main advantages over the standard level set method. First, new holes can be automatically generated in the design domain during the optimization process. Second, the dependency of the obtained optimized configurations upon the initial configurations is eliminated. Optimized configurations can be obtained even being started from a minimum possible initial guess. Third, the method can be easily implemented and is computationally more efficient. The validity of the proposed method is tested on the mean compliance minimization problem and the compliant mechanisms topology optimization problem.

Keywords:

• mean compliance
• compliant mechanisms
• topology optimization
• level set method
• bi-directional evolutionary algorithm

Acknowledgements

This research was supported by the National Natural Science Foundation of China (Grant No. 91223201), the Natural Science Foundation of Guangdong Province (Grant No. S2013030013355), Project GDUPS (2010), and the Fundamental Research Funds for the Central Universities (2012ZP0004). This support is greatly appreciated.