A two-stage stochastic model for co-firing biomass supply chain networks

A sustainable alternative to fossil fuels is biomass. The design of cost-efficient biomass networks is necessary to compete with nonrenewable resources. The creation of integrated biomass supply chain (BSC) network design models and solution procedures can contribute to the achievement of this goal. In particular, the emerging bioenergy industry must take advantage of the economies of scale in transportation to minimize the final product's cost. Hub-and-spoke networks have been proposed as a modeling approach to design large scale BSCs. The majority of these models are deterministic and do not consider the inherent variability in the biomass feedstock, such as physical and chemical properties that affect transportation, the effects of future climate on the biomass supply, overall cost, and production operations. Levels of ash and moisture are directly related to the quality of the feedstock, which negatively affects production of biofuels, increasing transportation and handling costs as well as the burden on the BSC's efficiency. Varying weather affects biomass yield, which creates a fluctuation in the incoming supply into the network, creating a complex large scale Newsvendors Problem. In this paper, a stochastic hub-and-spoke network model and an efficient solution are proposed to minimize logistical costs by finding an optimal production and distribution network as well as optimal transportation while reducing computational burden by using metaheuristics, such as Simulated Annealing, when solving large NP-hard instances. The proposed solution procedure was compared to recent benchmark models, i.e. Bender's Decomposition, yielding an overall solution difference of less than 7.71% and average reduction time difference of 91.4%. Case studies with several scenarios based on varying weather conditions were created using realistic data from the northeast region of the U.S.