2018 Optimization Days
HEC Montréal, Québec, Canada, 7 — 9 May 2018
TB6 Supply chain I
May 8, 2018 03:30 PM – 05:10 PM
Location: Serge-Saucier (48)
4 Presentations
-
03:30 PM - 03:55 PM
Formulations and branch-and-cut algorithms for multi-trip drone routing problem with time windows
This paper studies a multi-trip drone routing problem with time windows, where drones' energy consumption is influenced by payload and travel distance. To tackle the nonlinear energy function, we propose two types of cuts to calculate it exactly, instead of using linear approximation methods. Two formulations are presented, which are solved by branch-and-cut algorithms. Benchmark instances are first generated for this problem. Numerical tests evaluate the performance of the formulations and compare the solutions generated by exact and approximate methods.
-
03:55 PM - 04:20 PM
Order picking in e-commerce warehouses
We tackle a problem which is closely related to the Amazon’s PrimeNow service that proposes its customers to deliver them in a competitive time-lag (1 H after the demand’s release date). We focus on the order picking process. We aim to organize the order picker’s routes to satisfy all the demands and minimize the order pickers number. We propose a MILP model and a two-stage heuristic to solve the problem.
-
04:20 PM - 04:45 PM
An online stochastic optimization approach for the management of "smart warehouses"
Robot managed ‘smart warehouses’ are winning popularity as with online retailers such as Amazon and Alibaba. We consider the joint problem of requests sequencing and storage allocation using a fleet of robots. To account for the dynamic nature of the problem, a subset of requests is revealed through the day. This study proposes an online stochastic optimization framework for this specific version of an Automated Storage and Retrieval System.
-
04:45 PM - 05:10 PM
Warehouse task prioritization
Warehouse management involves a set of diverse problems. One of them concerns the prioritization of picking tasks to employees to minimize global treatment time while taking into account the warehouse state. We show how task management is usually designed and different alternative approaches that we experimented at JDA labs.