An Improvement of a Heuristic Algorithm for 3D Bin-Packing Problem

We are given the large 3-dimensional rectangular items, which we refer to as the bins, in the 3D bin-packing problem. They have specific weight, length, height, and depth measurements. Additionally, we have various small, rectangular, 3-dimensional items that we call boxes. These boxes also have specific weights, lengths, heights, and depths. The goal of this problem is to pack all the small boxes into the bins so that their combined weight does not exceed the weight limit of each bin. Additionally, we aim to use as few bins as possible. In this paper, we refine the computational processes to enhance an algorithm by Youssef Harrath [1]. Experimental results indicate that as the number of boxes increases, our improvements can reduce the number of bins used, decrease running time, increase efficiency, and reduce memory usage. That is, the number of bins used, processing time, and memory use were all decreased to 38.68%, 58.42%, and 63.33%, respectively. In the meantime, bin space utilization efficiency scaled to 35.38%.