Mixed integer non-linear programming using cutting plane techniques

In the present paper a modification of the extended cutting plane (ECP) method is described and illustrated. It is shown how it is possible to solve general MINLP (Mixed Integer Non-Linear Programming) problems with pseudo-convex objective as well as constraints to global optimality by a sophisticated cutting plane approach. The method relies on the ability to construct valid cutting planes for the entire feasible region of the problem. The method is illustrated on a simple test example and on some demanding practical scheduling problems. A comparison with a recently developped branch-and-bound approach is also given.