We propose a new algorithm for solving integer programming (IP) problems th
at is based on ideas from algebraic geometry. The method provides a natural
generalization of the Farkas lemma for IP, leads to a way of performing se
nsitivity analysis, offers a systematic enumeration of all feasible solutio
ns, and gives structural information of the feasible set of a given IP. We
provide several examples that offer insights on the algorithm and its prope
rties.