Green Order Orchestration for E-commerce Deliveries
Résumé
E-commerce has grown impressively the last few years. Unfortunately, so have the environmental impacts of its logistics. One of the key challenges in retail logistics is the order orchestration, which corresponds to the dispatch of online orders to stock locations. In general, this process relies on heuristic rules. To improve it and reduce environmental impacts, we propose in this paper two mathematical optimisation models to orchestrate orders while minimising the environmental impacts. Contrary to the model proposed in [21], we assume that items are non-breakable. Using real world and realistic datasets, we perform experiments to compare our models and the one of [21] on computational performances and environmental relevance. Then, we validate the relevance of our comprehensive environmental objective function by comparing our model with a parcels distance minimisation problem, showing that this simpler criteria increases environmental costs by 15%. A second comparison is performed with an economic criteria, revealing that favouring the environmental criteria only increases economic costs by 10%, while favouring economic costs increases environmental costs by more than 50%. Finally, we compare our results with the historical orchestration of the real world data that includes operational constraints that are not considered in our model. It highlights that our model is better on both environmental and economic aspects. The results of our paper should encourage retailers to consider mathematical optimisation for their order orchestration and to investigate the benefit of certain operational constraints.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| Licence |