Multi-Scale Optimization of Biomass-to-Bioenergy Supply Chains

This thesis deals with the development of discrete-continuous optimization models and algorithms that address multi-scale optimization of biomass-to-bioenergy supply chains. Specifically, contributions to the following three areas are presented, namely sustainable supply chain optimization, game-theoretic optimization for decentralized supply chains, and optimization under multi-scale uncertainties.;Sustainable supply chain optimization aims to quantitatively evaluate and systematically improve the sustainability of biomass-to-bioenergy supply chains. In the first related project, we improve the classical life cycle optimization (LCO) framework by introducing the theory of functional-unit-based optimization and two tailored optimization algorithms for efficiently solving the resulting mixed-integer linear fractional programming problems. In the second related project, we propose a hybrid LCO framework, which overcomes the drawbacks of commonly used process-based LCO, namely system boundary truncation and underestimation of the true impact. The hybrid LCO framework quantifies both direct and indirect environmental impacts and incorporates them into the decision-making process.;Unlike a conventional centralized approach, game-theoretic optimization for decentralized supply chains assumes that supply chain participants may have their own decision variables and objective functions, and very often, conflicts of interest. In the first related project, we employ the Nash bargaining solution approach for fair profit allocation among cooperative supply chain participants by optimizing material transfer prices and revenue sharing policies. In the second related project, we employ a single-leader-multi-follower Stackelberg game and the generalized Nash equilibrium to model competition among non-cooperative supply chain participants. In the third related project, we resolve the computational challenges caused by the incorporation of a follower's discrete decisions in the Stackelberg game and propose a novel reformulation-and-decomposition algorithm for mixed-integer bilevel linear programs.;In the project of optimization on multi-scale uncertainties, we propose a stochastic robust optimization model and a decomposition-based algorithm, aiming to optimize the expected economic performance while ensuring the robustness of operations. Unlike a typical "one-size-fits-all" approach, we distinguish between strategic and operational uncertainties and allow the decision maker to have different levels of conservativeness towards different uncertainties.