K-adaptability in robust container vessel sequencing problem with week-dependent demands of a service route

This work investigates a robust container vessel sequencing (RCVS) problem in a service route. As weekly demands vary dramatically and cannot be forecasted accurately, shipping companies need to develop a robust sequence of vessels with different capacities to maximally meet demands. As export heavily depends on economy, demands may share the same pattern in adjacent years, which motivates us to study the problem in a cyclic fashion. To refine the literature, we adopt a robust optimisation model to minimise the worst-case total cost, including container tardy and outsourcing cost, due to reliability guarantee. To accommodate human decision-making, we focus on an associated K-adaptability problem, which pre-selects a number of candidate vessel sequences and implements the best one when the uncertain demands have been observed. A branch-and-bound solution approach is explored. Numerical experiments demonstrate the performance of our approach.