We address the Thief Orienteering Problem (ThOP), a multi-component problem that combines features of two classic combinatorial optimization problems, namely the Orienteering Problem and Knapsack Problem. Due to the given time constraint and the interaction of the load-dependent movement speed with the chosen route, the ThOP is complex and challenging. We propose a two-phase, swarm-intelligence based approach together with a new randomized packing heuristic. To identify the impact of the respective components, we use automated algorithm configuration. The resulting configurations outperform existing work on more than 90% of the benchmarking instances, with an average improvement of over 300%.