A Fast Optimization Approach For A Complex Real-Life 3D Multiple Bin Size Bin Packing Problem

TL;DR

This paper presents a fast, heuristic-based optimization method for a complex 3D bin packing problem involving real-life air cargo constraints, improving solution efficiency and practicality.

Contribution

It introduces an extended extreme point concept, a layered sorting strategy, and space division techniques to enhance packing solutions for complex cargo loading.

Findings

Effective on real-life cargo instances

Improves packing efficiency and stability

Reduces computational time for complex constraints

Abstract

We investigate a real-life air cargo loading problem which is a variant of the three-dimensional Variable Size Bin Packing Problem with special bin forms of cuboid and non-cuboid unit load devices (ULDs). Packing is constrained by additional practical restrictions, such as load stability, (non-)stackable items, and weight distribution constraints. To solve the problem, we present an insertion heuristic embedded into a Randomized Greedy Search. The solution space is limited by only considering certain candidate points (so-called extreme points), which are promising positions to load an item. We extend the concept of extreme points proposed in the literature and allow moving extreme points for non-cuboid ULDs. A special sorting of the items is suggested, which combines a layered structure and free packing. Moreover, we propose dividing the space of each ULD into smaller cells to accelerate the collision, non-floating, and stackability check while loading items. In a computational study, we analyze individual algorithm components and show the effectiveness of our method on adapted real-life instances from the literature.