Column generation approaches to large driver scheduling problems

The integer linear programming (ILP) approach plays an important role in scheduling public transport drivers, which needs heuristics to reduce the size of problems. Optimality is therefore compromised. This paper presents a column generation technique and the revised simplex approach, which can enhance the ILP solver to solve larger problem instances. Two different column generation strategies for driver scheduling were analyzed based on the well-known driver scheduling approaches: TRACS II and HASTUS, whilst the definition of the reduced cost for non-basic variables was depicted. The limitations of the two successful approaches were finally summarized, from which some further research could be found.