Biofuels are currently the only demonstrated replacement for liquid transportation fuel, owing largely to the expansion of the U.S. and Brazilian bioethanol industries. Biofuels have the potential to reduce petroleum imports, support domestic agriculture and forestry, and reduce greenhouse gas emissions relative to gasoline. In the U.S., the Renewable Fuels Standard (RFS) Program phases in the use of biofuels by setting annual blending mandates for renewable, advanced and cellulosic biofuels. The mandates for renewable fuel have so far been met largely by corn-based bioethanol, but cellulosic biofuel is lagging behind. For 2011, the original cellulosic biofuel mandate of 250 million ethanol-equivalent gallons was reduced by the EPA to 6 million gallons based on expected industry production. Tools are required to analyze how the industry may develop to produce the 16 billion gallons of cellulosic biofuels required by RFS in 2022. This study aims to address this issue by using computational optimization to generate robust supply chains using the real-world distribution of biomass resources, existing options for biomass collection and transportation, and proven biomass-to-biofuels technologies. The Midwest is chosen as the study region due to the abundance of biomass there and its expected importance in fulfilling the RFS mandates. Forestry and agricultural residues are included in the optimization as available biomass resources with county-level resolution. Two cellulosic biofuel technologies are allowed to compete in the study region, including co-fermentation (biochemical) and gasification Fischer-Tropsch (thermochemical). The problem is formulated as a multi-objective mixed integer linear program (MoMILP). The competing objectives are to maximize the net present value (NPV) and minimize the total greenhouse gas emissions of the whole supply chain by taking into account cash flows and emissions for biomass production, collection and transportation, biorefinery construction and operation, and biofuel sales. The effect of uncertainty in biomass availability and in economic parameters is examined by Monte Carlo sampling.