EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

2079. Joint schedule and volume optimization at scale

Invited abstract in session WB-57: Scheduling and sustainability, stream Optimization at Amazon.

Wednesday, 10:30-12:00
Room: S06 (building: 101)

Authors (first author is the speaker)

1. Marc Bataillou Almagro
Amazon
2. Konstantinos Benidis
Amazon
3. Martin Gross
Amazon

Abstract

The joint schedule and volume optimization aims at creating an optimal truck schedule with a corresponding volume plan that can satisfy the demand of the network in a speedy manner and with minimum cost. The speed of delivery is related to the path that a package follows from an origin up to a final destination node, while the cost is proportional to the number of trucks that we schedule in order to move all the required packages. The problem has many operational and practical constraints that make it challenging. For example, all sites have inbound and outbound restrictions and capacity constraints, the packages are consolidated in intermediate nodes (sort centers) in various ways, there are driving bans that the schedule needs to comply with, etc. The problem is formulated as an MILP, to which we apply a set of techniques to reduce complexity such as smart pruning of decision variables and solving sequentially the problem to incrementally finer granularities. Experimental results on large-scale networks illustrate the effectiveness and scalability of our method.

Keywords

Status: accepted

Back to the list of papers