An intelligent framework to online bin packing in a just-in-time environment

Abstract

This paper addresses a complex real life manufacturing problem that occurs in paper and wood industries: the online bin packing and cutting of small items using parallel machines in a just-in-time environment. The objective is to minimize both the unused areas of the bins and the sum of weighted earliness and tardiness penalties. This NP hard problem is herein solved using an agent-based approach A-B. The active agents of A-B interact dynamically in real time to jointly fill the bins with items, assign cutting patterns to machines, cut them, and have the items delivered on time. The application of A-B to real life instances from a manufacturing company highlights the sizeable savings induced by A-B, and demonstrates its suitability for complex online real-life situations.