The Lens Design Using the CMA-ES Algorithm

Abstract

AbstractThis paper presents a lens system design algorithm using the covariance matrix adaptation evolution strategy (CMA-ES), which is one of the most powerful self-adaptation mechanisms. The lens design problem is a very difficult optimization problem because the typical search space is a complicated multidimensional space including many local optima, non-linearities, and strongly correlated parameters. There have been several applications of Evolution Algorithms (EA) to lens system designs, and Genetic Algorithms (GAs) are generally expected to be more suitable for these kind of difficult optimization problems than Evolution Strategy(ES) because GAs can provide a global optimization. We demonstrate, however, that a CMA-ES can work better than the GA methods previously applied to lens design problems. Experimental results show that the proposed method can find human-competitive lens systems efficiently.KeywordsSearch PointLens SystemParaxial ApproximationCovariance Matrix Adaptation Evolution StrategyLens DesignThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.