Determining a Geographic High Resolution Supply Chain Network for a Large Scale Biofuel Industry
Lambert, L. H., B. C. English, D. M. Lambert, O. Shylo, J. A. Larson, T. E. Yu, and B. S. Wilson.  2018.  Applied Energy, 218: 266-281.

This research combines Geographic Information System (GIS) and Mixed Integer Programming (MIP) models to determine feedstock supply and facility locations at high-resolution spatial units to meet large-scale annual biofuel production and demand goal in the state of Tennessee. Because the large amount of integer variables and continues variables associated with the MIP model, the challenge exists if all the variables to be determined simultaneously. The solver routine cannot be read into a typical personal computer or can be read into High Performance Computer, but could not determine the feasible solution in 96 hours. This paper propose a two-step optimization procedure to overcome these computational challenges while keeping consistent with the firm location theory. Results indicate that with the two-step optimization approach, the solution of the entire state of Tennessee region can be obtained in less than a day. The comparison of solution using a two-steps and a simultaneous location model approach were also provided using a sub-regional demonstration model.