Hybrid metaheuristics for packing problems

Abstract

Three variants of the two dimensional packing problem are considered, where the items to be packed are (a) rectangles with fixed widths and heights, (b) rectangles with adjustable widths and heights, or (c) irregular shapes. All problems are solved by hybrid metaheuristics that combine local search and mathematical programming techniques of linear, nonlinear and/or dynamic programming. Basic ideas of these algorithms are explained on a unified basis, together with some computational results. It appears to indicate that mathematical programming is a vital tool for enhancing metaheuristic algorithms. © 2008 Springer-Verlag Berlin Heidelberg.