Abstract
Progress in global optimization methods and the availability of fast computers have lead in the past few years to increasingly powerful lens design software that is capable to detect multiple local minima corresponding to given system specifications. However, there is still no known technique to determine whether the local minimum having the lowest value of the error function among those already detected is indeed the global minimum, especially if the number of optimization variables is large. While this continues to be a very difficult problem, in this paper we present a new technique of optical system analysis which may shed some new light on this issue.