Simulation Framework for Bin Packing ProblemsTri Thanh- Le
A framework is developed to generate benchmark problems for dynamic three-dimensional (3D) bin packing problems (BPPs). This framework is able to generate benchmark problems for different variants of BPPs by taking into account potential uncertainty in real-world BPPs, which are uncertainties in dimensions, costs, weights of upcoming items. This paper has three main contributions.
First, a benchmark generator framework is developed for the first time using an open source discrete-event simulation platform. This framework generates benchmark problems for BPPs by reproducing uncertainty in real-world BPPs. Second, this framework can be integrated with any dynamic BPP algorithm so that the optimisation algorithm can be run alongside the simulation to solve BPPs under uncertainty. Third, various performance measures from the literature are included in the framework to evaluate the optimisation algorithms from different perspectives.
Thanks to the 3D visualisation feature of this framework, the optimisation results can also be observed visually. Finally, empirical experiments on a real-world BPP are conducted to verify these contributions.
This work was supported by a British Council UK-ASEAN Knowledge Partnership grant and a British Council Newton Institutional Links grant.
More detail at http://ukvn-logistics.org/ran-wang-trung-thanh-nguyen-shayan-kavakeb-zaili-yang-changhe-li-2016-benchmarking-dynamic-three-dimensional-bin-packing-problems-using-discrete-event-simulation-european-leading-conference-evo/