Structural design of a lens component with prespecified aberration targets by evolutionary algorithm

Abstract

In a recent paper we have reported on the feasibility of reducing the problem of structural design of a multicomponent lens system in accordance with a prespecified set of Gaussian characteristics and primary aberration targets to the relatively simpler problem of determining optimum structures for the component lenses with central aberration targets. The structural design of the individual components can be suitably tackled by evolutionary algorithm so that one can obtain globally or quasiglobally optimum solutions for the purpose. Details of the latter part of the approach are presented in this paper. Contrary to the usual practice of heuristic selection of glass types for the lens elements of the component, the glass types for individual lens elements are treated as discrete independent variables to be selected from a set of prespecified list of actual glasses. We have dealt with the global or quasiglobal synthesis of the individual lens components with the help of structures of increasing complexity, e.g. singlets, cemented doublets, broken contact doublets, cemented triplets, photovisual objectives etc. as necessary for the purpose. The total configuration space consists of continuous variables like shape variable and power distributions, and discrete variables like available glass types. This approach reduces significantly the chance of overlooking promising and better solutions by carrying out searches in the total configuration space simultaneously. Some illustrative examples will be presented.