Generalized branch-and-cut framework for mixed-integer nonlinear optimization problems

Branch and bound (BB) is the primary deterministic approach that has been a pplied successfully to solve mixed-integer nonlinear programming (MINLPs) p roblems in which the participating functions are nonconvex. Recently, a dec omposition algorithm was proposed to solve nonconvex MINLPs. In this work, a generalized branch and cut (GBC) algorithm is proposed and it is shown th at both decomposition and BE algorithms are specific instances of the GBC a lgorithm with a certain set of heuristics. This provides a unified framewor k for comparing BE and decomposition algorithms. Finally, a set of heuristi cs which may be potentially more efficient computationally compared to all currently available deterministic algorithms is presented. (C) 2000 Elsevie r Science Ltd. All rights reserved.