Fairness and efficiency in railway freight transport
Many goods in Europe are still transported by railway. In order for that to function, train driver satisfaction and motivation is crucial – efficiency and fairness must be taken into account when scheduling railway crews. Prof Dr Silke Jütte of IUBH University of Applied Sciences conducted research on this topic.
The bulk of railway freight traffic in Europe is handled at night. This means that many work shifts for train drivers are at night and in the early morning. Work shifts that start in the early morning hours between two and five are particularly unpopular with railway crew because it has been shown that nocturnal work alters biorhythms and throws the inner biological clock out of kilter. If such unpopular work shifts cannot be avoided, they can at least need be distributed as evenly as possible among the various crew depots – otherwise there is the danger of low employee motivation or even strikes.
Levers for more fairness
My colleagues and I therefore examined how aspects of fairness could be included in the existing scheduling planning process of a large European railway freight service provider. We integrated two additional levers into the computer controlled mathematical optimisation process: one with which the number of unpopular shifts in the schedule can be reduced, and one in which the existing unpopular shifts are distributed as evenly as possible among the individual locations in proportion to the number of employees. Both levers are, however, associated with additional operational costs: in each case, schedules with lower productivity than before must be considered.
For example, suppose a schedule has two unpopular shifts that start at two o’clock in the morning. Currently, both shifts are assigned to the crew depot in Cologne. In contrast, for train drivers at the second crew depot in Frankfurt, no unpopular shifts have been assigned. This unfair situation could be resolved in a variety of ways:
- Change the schedule so that it no longer includes unpopular shifts. To do this, the shift that used to start at two a.m. must be combined with duties that preceded them in time. In general, the changed schedule is less productive because, for example, train drivers must accept longer waiting times.
- Change the schedule so that one of the unpopular shifts is assigned to the Frankfurt location. This would mean, for example, that a train driver must take a passenger train from Frankfurt to Cologne before he/she can drive the freight train. Such a change in the schedule plan is therefore associated with additional costs.
- Change the schedule so that two additional unpopular shifts are assigned to Frankfurt. The schedule now has a total of four unpopular shifts (before, it was two), but these are evenly distributed among the crew depots.
Compromises are unavoidable
The example illustrates the various possibilities for creating a schedule with different goals. If one goal is reached (such as a schedule with increased fairness), one generally has to accept that other goals are compromised (such as additional costs associated with the schedule). The compromises made depend on the respective corporate strategy.
Further information can be found in the Journal of Scheduling 2017, Volume 20, Issue 1, pp 43-55: “Optimizing railway crew schedules with fairness preferences” by Silke Jütte, Daniel Müller, Ulrich W. Thonemann