Optimized Tugboat Scheduling in the Port of Rotterdam

In this thesis we investigate optimized scheduling of tugboats in the port of Rotterdam. Three heuristic methods for generating schedules, based on Tabu Search, Ant Colony Optimization (ACO) and Genetic Algorithms (GA), are compared with respect to solution quality and suitability for application. Optimal results computed by solving an integer programming problem serve as a benchmark for the heuristic algorithms. Practical requirements for tugboat scheduling include delivering solutions in $60$ seconds, putting restraints on each algorithm. We test the algorithms on real life scenarios from the port of Rotterdam. The key performance indicators for the algorithms are the total cost over all scenarios and the number of times the algorithm failed to produce an applicable solution within $60$ seconds. The algorithm based on Tabu Search outperforms the other two algorithms with respect to both total cost and failure rate. This algorithm is best suitable for real life application, though we also flag possible areas of improvement.