A cut-map algorithm for design problems with parameter tolerances

The design problem of choosing a set of parameters so that inequality constraints are satisfied for a specified variation of parameter values about their nominal value is considered. Such problems occur when systems must be synthesized from components whose values are only known to a certain tolerance. Simple algorithms exist for such problems when the constraints are convex. This paper presents an algorithm which is valid for the nonconvex case. The algorithm utilizes concepts employed by Eaves and Zangwill in their generalized cutting plane algorithms.