Dynamic Planning of Facility Locations with Benefits from Multitype Facility Colocation

This paper presents a mixed-integer programming model for a multi-type facility co-location problem with capacity expansion over a multi-period horizon. This problem is motivated by the emerging U.S. biofuel industry that is rapidly expanding its infrastructure facilities to produce bioethanol from agricultural crops. Each type of facility (e.g., corn- or cellulose-based bio-refinery) requires a different type of raw material (e.g., corn or perennial grass), but they produce the same final product (e.g., bio-ethanol), and co-location of multiple types of facilities results in cost-saving benefits due to complementary production process and byproduct recycling. Multiple solution approaches (i.e., Lagrangian relaxation, Benders decomposition, and accelerated Benders decomposition) are proposed to solve this problem. Numerical experiments show that accelerated Benders decomposition most effectively solves large-sized problems in a short amount of time. Various managerial insights are also drawn from the computational results.