Abstract
In this paper, we describe a general non-simplex type algorithm for determining all essential solutions and inequalities for any convex polyhedron defined either in a tangential or in a barycentric form. This algorithm is based on Motzkin's original Double Description Method, with some additional redundancy criteria in order to eliminate all the extraneous solutions and inequalities. These criteria are compared with other redundancy criteria published in the literature. Also, some simple examples and counter-examples highlighting their scope of validity are exhibited. The theoretical validity of the entire algorithm is proved in a straightforward way using Uzawa's algorithm. As a consequence, we obtain a constructive proof for the existence of an essential double description for any convex polyhedron. Finally, some classical feasibility conditions for general linear inequality systems are recovered.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
