Liner shipping single service design problem with arrival time service levels

We introduce three mathematical models of increasing complexity for designing liner shipping services that guarantee the punctual arrival of vessels at a specified service level. On-time reliability is an important performance indicator for many liner carriers, but current approaches for creating new routes in liner shipping networks do not consider data-driven uncertainty. We perform an empirical analysis of vessel travel times in a real liner shipping network to develop probability distributions that we use within novel, chance-constrained mathematical models for liner shipping service design. Our models are also the first to support variable vessel speeds for service design. In our experiments, we use real-world data from 22 liner shipping routes and evaluate the designed services using a simulation procedure that demonstrates the effectiveness of our approach for reducing lateness. We show that our models can be effectively used for decision support at a tactical level not only for designing services, but also potentially for negotiating maximum demand transit times and prices with customers.