Comparing two optimization approaches for ship weather routing
Presentation held at OR 2016, International Conference on Operations Research of the German Operations Research Society (GOR), Hamburg, Aug 30 - Sept 2, 2016
Weather routing in maritime shipping is related to a shipping company's objective to achieve maximum efficiency, economy and cost competitiveness by optimizing each voyage of a ship. A voyage can be optimized regarding costs, time, safety or combinations of these, while taking into account forecasted meteorological and oceanographic information as well as constraints given by geographic conditions, ship characteristics, emission regulations, safety requirements or time restrictions. A wide variety of mathematical models of the ship weather routing problem can be found in the literature. The formulations vary from constrained graph problems to nonlinear optimization problems and from one objective to multiple objective optimization problems. Numerous commercial systems and academic developments apply different approaches for solving the optimization problem, which range from calculus of variations, dynamic programming or discrete optimization methods to evolutionary methods. In this paper two ways to approach the ship weather routing problem using a discrete optimization method on the one hand and an evolutionary algorithm on the other are presented. For both models, the ship's heading and its delivered engine power are introduced as control variables to allow route and speed optimization. The two approaches aiming at minimum fuel costs are compared based on numerical examples with real-world data.