A Characterization of Ill-Posed Data Instances for Convex Programming

Manuel A. Nunez

Given a data instance of a convex program, we provide a collection of conic linear systems such that the data instance is ill-posed if and only if at least one of those systems is satisfied. This collection of conic linear systems is derived from a characterization of the boundary of the set of primal and dual feasible data instances associated with the given convex program.

School of Business and Economics, Chapman University, 333 N Glassell St., Orange, CA 92866, September 10, 1998.

Contact: mnunez@chapman.edu