Seminer duyurusu (5 Aralık 2014)

SEMINAR 

DEPARTMENT OF INDUSTRIAL ENGINEERING 

An Online Optimization Approach for Outbound Logistics

by

Burcu Keskin,

(University of Alabama, Operations Management

 

Abstract:

We present an integrated model for simultaneous optimization of the loading and routing decisions associated with an automotive supplier's outbound supply chain. The supplier observes and fulfills demand in an online fashion.  Given limited inventory, the supplier combines these orders into loads, releases them based on various dispatch criteria (e.g. truck-utilization, route-utilization, or penalty-based dispatch policies) and due-date considerations, and ships them in truckloads, less-than-truckloads, and containers. Dynamically changing demand information, inventory rationing, inventory interactions among orders, and lead-time considerations are some of the challenging aspects of the problem. Considering both a decomposition-based heuristic and a reduction-based exact approach, we determine transportation mode and carrier choices, ration inventory, and route orders while considering total cost minimization. The heuristic approach decomposes the problem into assignment and routing components following a cluster first-route second logic. On the other hand, due to the newly developed problem size reduction techniques, the exact approach is solved via an off-the-shelf optimization engine. A scalable database with a graphical user interface supports the optimization algorithms. We test our algorithms using four months of real data from the supplier and compare it to their real life practice. Our results demonstrate the impact of transportation mode-specific capacities, customer locations, inventory availabilities, and due- date restrictions on outbound logistics costs. 

  

Short Bio: 

We present an integrated model for simultaneous optimization of the loading and routing decisions associated with an automotive supplier's outbound supply chain. The supplier observes and fulfills demand in an online fashion.  Given limited inventory, the supplier combines these orders into loads, releases them based on various dispatch criteria (e.g. truck-utilization, route-utilization, or penalty-based dispatch policies) and due-date considerations, and ships them in truckloads, less-than-truckloads, and containers. Dynamically changing demand information, inventory rationing, inventory interactions among orders, and lead-time considerations are some of the challenging aspects of the problem. Considering both a decomposition-based heuristic and a reduction-based exact approach, we determine transportation mode and carrier choices, ration inventory, and route orders while considering total cost minimization. The heuristic approach decomposes the problem into assignment and routing components following a cluster first-route second logic. On the other hand, due to the newly developed problem size reduction techniques, the exact approach is solved via an off-the-shelf optimization engine. A scalable database with a graphical user interface supports the optimization algorithms. We test our algorithms using four months of real data from the supplier and compare it to their real life practice. Our results demonstrate the impact of transportation mode-specific capacities, customer locations, inventory availabilities, and due- date restrictions on outbound logistics costs. 

 

All interested are cordially invited.  

 

DATE:  December 5, 2014 

TIME:  Friday, 15:00-16:00 

ROOM: VYKM 3