Clustering Search and Simulated Annealing to solve the driver scheduling problem for timber transport.

The transportation of cargo in Brazilian forestry is mostly done by roads. Currently the driver scheduling problem has been a major obstacle to optimize activities in transport companies, due to severe physical and technical constraints. This research aimed to apply two meta-heuristics (Clustering Search -CS and Simulated Annealing -SA) to solve problems of driver scheduling for timber transportation in a forestry company located at Vale do Rio Doce MG. We used input generated from a qualitative research and time and motion studies to validate the proposed approach, measuring the travel times of the vehicles and defining the points where driver exchanges are done. Computational results indicate that the methods (CS and SA) succeeded in reducing the150 drivers by 1/3 and in eliminating all of the daily overtimes and extra hours.