An Efficient Algorithm for Solving the Container Loading Problem

Abstract

We present a new algorithm for the three-dimensional packing problem with a single container to be loaded. Deviates from the traditional approaches, it uses a principle —— “largest caving degree” such that items are packed as closely as possible. Then, it incorporates a backtracking strategy that gains a good trade-off between solution quality and computation time. We tested our algorithm using all the 47 related benchmarks from the OR-Library that have no orientation constraints. Experiments indicate the algorithm’s good quality. The container volume utilization, achieved within comparable time, is 94.6% on average. This significantly improves current best-known results in the literature by 3.6%.