Ferry service network design under demand uncertainty

• A new service reliability (SR)-based service network design formulation is developed.
• The formulation address demand uncertainty via recourse consideration.
• A corresponding SR-based gradient method is developed to solve this formulation.
• We prove that its solutions are consistent with existing 2-stage stochastic programs.
• We demonstrate the approach for ferry service network design in Hong Kong.

This paper develops a formulation for ferry service network design with stochastic demand via the notion of service reliability (SR). The problem is formulated as a two-phase stochastic program in which the schedule of regular and ad hoc services are derived sequentially. An example is presented to illustrate the formulation. We then apply the proposed method to the case of ferry network design in Hong Kong. In terms of solution quality and computational efficiency, the SR-based approach shows promising results as compared with existing methods. Moreover, the approach leads to substantial cost savings as compared with deterministic methods under demand uncertainty.