Accounting for demand uncertainty when designing the network of express integrators
(2021) IFORS 2021 Virtual — Location: Online (23.August.2021)
Files
No attached file found for this publication.
Details
- Authors
- Abstract
- The Express Shipment Service Network Design (ESSND) problem consists in determining a schedule of daily flights to efficiently deliver packages in large regions overnight. Even though express integrators face demand variability from day to day, the ESSND literature only considers deterministic demands. In this work, we account for demand uncertainty and aim to balance service level and cost when designing the network. We propose the aircraft routing model with demand uncertainty (ARMdu), which builds on composite variables to account for demand uncertainty. Thanks to their small size, the service level of the composite variables can be computed efficiently using simulation. Then, the network service level can be estimated as a function of the selected composite variables. The ARMdu maximizes this estimation of the network service level while limiting the network cost. An alternative model, minimizing the cost to reach a given service level, is also formulated. The effectiveness of our approach is shown on instances of realistic size provided by an express integrator in Europe. We show that, compared to the deterministic approach used in practice, the ARMdu allows to significantly improve either the network service level (number of undelivered packages reduced by 24%) or the network cost (reduced by 1.1%), and it does so in a short computational time (around three minutes).
- Affiliations
Citations
Tancrez, J.-S., Quesada Pérez, J. M., & Lange, J.-C. (2021). Accounting for demand uncertainty when designing the network of express integrators. IFORS 2021 Virtual, Online. https://hdl.handle.net/2078.5/114105