Theses and Dissertations



Marufuzzaman, Mohammad

Committee Member

Bian, Linkan

Committee Member

Wang, Haifeng

Committee Member

Gnaneswar Gude, Veera

Date of Degree


Original embargo terms

Visible MSU only 6 month

Document Type

Dissertation - Campus Access Only


Industrial Engineering

Degree Name

Doctor of Philosophy (Ph.D)


James Worth Bagley College of Engineering


Department of Industrial and Systems Engineering


Biomass-based CHP (bCHP) can provide reliable electricity in remote and rural areas because it is an on-site generation resource, and it is designed to support continued operations in the event of a disaster. However, the benefits of such facilities can only be realized if a reliable and economical feedstock supply system is designed, given the system not only efficiently transports biomass under normal scenarios (e.g., when depots and transportation links are functioning properly) but also hedges against unexpected infrastructure/transportation link failures due to severe weather events (e.g., hurricanes). To serve this purpose, this study proposes a three-stage stochastic programming model to design a reliable feedstock supply system, where decisions are made sequentially to realistically represent pre-and-post disaster situations) under uncertain infrastructure status (e.g., unavailability of the road and facility conditions) and customer demand situations. In stage one, pre-disaster decisions are made (e.g., the opening of depots and regular feedstock transportation decisions), while stages two and three represent, respectively, immediate decisions following a disaster (e.g., damaged timber transportation, pellet production) and post-disaster decisions (e.g., transportation pellets to end-users, storage) with a timeframe between several days to weeks. By collecting data from 15 coastal rural counties in Mississippi, we create a real-life case study and derive important managerial insights. Our experimental results reveal that the biomass-to-bCHP supply chain decisions (e.g., depot location, storage, transportation decisions) are highly sensitive to intensity and the probabilistic infrastructure availability following a hurricane. The second chapter extends the research by introducing high and low priority end-users so the demand prioritization is met.