The robust pickup and delivery problem with time windows

This dissertation introduces the robust pickup and delivery problem with time windows (RPDPTW), incorporating uncertainty in both demand and travel times through robust optimization theory. New mathematical formulations are proposed, including a multi-commodity flow model and a dynamic programming–based approach that explicitly exploits the pairing between pickup and delivery nodes. Specialised exact algorithms and valid inequalities are also developed to efficiently solve large-scale instances. Computational experiments show that robust solutions significantly reduce the risk of infeasibility with a moderate increase on routing cost, while also favouring routes with a higher degree of pairedness, with practical implications for variants with loading constraints and handling costs. The research resulted in a paper published in the European Journal of Operational Research.