Models and Algorithms for Distributed Order Management

Models and Algorithms for Plannning and Scheduling Problems (MAPSP) |

Author's Version

The emergence of cloud computing has revolutionized various business sectors, such as transportation, health care and retail. In particular, the ability to bring together huge amounts of data with accessible compute resources, has opened the gate for solving large-scale decision problems associated with the underlying data. One such example is Distributed Order Management (DOM), a key component in modern retail applications. The goal of DOM is assigning orders of customers to stores (or warehouses) while minimizing fulfillment costs (e.g., shipping or transport costs). We introduce here a formal mathematical framework to address the underlying optimization problems. In particular, we define several variants that differ in their objective (e.g., minimize fulfillment cost, maximize satisfied orders) and constraints (e.g., unsplittable vs. splittable assignments of orders to stores). We present approximation algorithms for the different models with proven lower and upper bounds.