The robust asset-liability management problem under return and interest rate uncertainty

Abstract

This thesis explores robust optimization theory and applications through a collection of articles. The first article reviews robust portfolio selection literature, identifying research gaps and paving the way for subsequent work. The second and third articles address asset liability management for pension funds under uncertain parameters like asset returns and interest rates. The second article introduces a distributionally-robust chance-constrained programming approach, while the third employs distributionally-robust optimization with various ambiguity sets. Both articles propose tractable reformulations and demonstrate enhanced asset allocation and funding ratio through realistic experiments. The final article focuses on the K-adaptability problem, presenting a solution method involving logic-based Benders decomposition. This approach, extended to cases with stage-wise uncertainty and nonlinear functions, outperforms existing methods. In essence, the thesis contributes valuable methodologies for optimizing in the face of uncertainty across different financial scenarios.