Lagrangian duality for 0-1 concave-convex frational programs

This paper deals with solving 0-1 concave-convex fractional programming pro blems (maximization of a ratio of a concave function over a convex function subject to convex constraints). For these models whose objective function is nonlinear and nonconcave, a new Lagrangian decomposition, based on a fra ctional rewriting of the constraints, is proposed and compared with the cla ssical Langranian relaxation. This Langranian decomposition combines solvin g a 0-1 constrained programming problem with linear function and solving a 0-1 unconstrained fractioning programming problem. (C) 2000 Academie des sc iences/Editions scientifiques et medicales Elsevier SAS.