A new sequential cluster algorithm for optical lens design (Q1093566)

Optical lens design problems are solved by numerical optimization. If we describe an optical lens system by constructive parameters, we may identify such a system by a vector \(x\in {\mathbb{R}}^ n\). To find an adequate system for a given problem, we have to organize the optimization process in a parameter space with a fixed dimension n. Often, such an optimization leads to inadequate results. For this reason, we have to alter the number of lenses contained in the system. Consequently, we have to change the dimension of the space for the parametric description of the lens system. This paper describes a special sequential cluster algorithm to realize the move from one parameter space to another by computation.