Abstract
We present an algorithm that certifies the feasibility of a linear program and computes a safe bound on its objective value while using rational arithmetic as little as possible. Our approach relies on computing a feasible solution that is as far as possible from satisfying an inequality at equality. To this end, we have to detect the set of inequalities that can only be satisfied at equality.Compared to previous approaches, our algorithm has a much higher success rate.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have