In order to solve these itinerary planning problems, Bock (2010) designs a variable neighborhood search algorithm and solves generated
instances with up to 5600 locations, 372 orders, 399 transshipment locations, and 55 vehicles. He shows that the exact
solutions are outperformed in a real-time environment and a continuous improvement results in substantial cost reductions. Goel
(2010) designs a simulation heuristic which at each iteration solves a multi-commodity network flow problem based on the updated
information. He compare the defined visibility levels on a small real-world network of 1 supplier, 3 rail operators with up to 40
trains per operator, and 3 factories. Goel (2010) concludes that on-time delivery performance can be significantly improved by
increasing the level of visibility.