The scheduling of transportation systems has traditionally been done once a day. At beginning of a working day, the planner establishes which tasks will be carried out by each vehicle. Then, traffic jam, breakdown and any unexpected problem will cause delays on our timetable. In this paper, we propose to use real-time vehicle position knowledge to solve this problem. So, the planner is permanently enabled to reallocate tasks as the problem conditions change. As both drayage problem is a NP-Hard problem and a high-speed procedure is required, exact methods are not computationally feasible. So, a genetic algorithm has been implemented to perform the problem described.