A cutting plane method for minimizing pseudo-convex functions in the mixed integer case

In this paper, a mixed integer non linear programming (MINLP) algorithm for minimizing pseudo-convex functions under pseudo-convex constraints is proposed and illustrated. The solution procedure is iterative and relies on successive linear approximation of the objective function and on a line-search technique. The whole procedure is then embedded within the framework of a existing cutting plane method for mixed integer non-linear programs. This enables us to solve general MINLPs with pseudo-convex objective and pseudo-convex inequality constraints to global optimality. (C) 2000 Elsevier Science Ltd. All rights reserved.