Abstract

This paper deals with the search for optimum solutions in optical designs. The design space of optical systems is typically a complicated multidimensional parameter space, which often has a large number of local minima. The limited successes of the conventional design method in finding the global minimum is due to the fact that the search in most optimization routines is crucially dependent on the initial configuration. This proposed search algorithm provides a set of promising initial configurations that can serve as an input to the second stage which is based on conventional optimization. First, the complete space is searched using a coarse interval for a collection of promising regions. This sampling procedure is enhanced by using parallel processing mechanisms. Specific heuristic optics rules are then applied to screen out some unfavorable configurations. Finally, after further optimization, a set of different configurations is constructed. This sampling algorithm was tested on some cases and was found to be an attractive new approach to optical design. It gives complete information on the performance limits, sensitivities, and trade-offs from a given design space. Furthermore, there is no need here to come up with any expert guesses for an initial configuration because of the systematic nature of this algorithm.

© 1989 Optical Society of America