Two-stage stochastic programming formulation for ship design optimisation under uncertainty

The paper presents a two-stage approach for ship design optimisation under uncertainties related to stochastic parameters that cannot be controlled by the designer. The designer decision is partitioned into two sets: the first contains variables that have to be decided before the onset of any uncertain condition, while the second Involves variables that may be decided once the random events have occurred. The first-stage variables are selected so that the sum of the first-stage cost and the expectation of the second-stage cost is minimised. The problem is solved as a two-nested-loop minimisation problem, which combines robustness of the first-stage decision with flexibility of the second-stage decision. The approach is applied to the conceptual optimisation of a bulk carrier. © 2010 Institute of Ship Technology and Ocean Engineering.