Date on Master's Thesis/Doctoral Dissertation


Document Type

Master's Thesis

Degree Name

M. Eng.


Industrial Engineering

Degree Program

JB Speed School of Engineering

Committee Chair

Bai, Lihui

Committee Co-Chair (if applicable)

Usher, John

Committee Member

Usher, John

Committee Member

Chang, Dar-Jen

Author's Keywords

Cost Allocation; Machine Learning; Transportation; Logistics


Advancements in big data enabled management practices inspire logistics companies to study deeper into their transportation operations with a data driven approach. One such question asks: How can a logistics firm identify high-cost customers in their service network? In the presence of rich data on routes involving many customers, this thesis develops a framework to allocate a route cost among customers that the route serves, where each route is associated with multiple route features related to the transportation cost. Cost is allocated using the proportional allocation approach in combination with the random forest method in machine learning. First, this framework ensembles random forest regression models to determine the importance values of all route features. Next, the importance values of route features are used to allocate cost among customers. Finally, posterior analysis identifies customers in a route or in general that are most costly to serve. Several additional analyses are performed to show potential uses of this cost allocation output. Results of the framework and analyses on three simulated case and two industry cases show the validity of the model and the potential for actionable operational analysis and changes.