This payer introduces an efficient adaptation of the branch-and-bound techn
ique that solves real-world rostering problems for airline crews. The effic
iency of the algorithm is based on the exploitation of rostering-specific p
roperties (e,g, variable selection, branching strategy and cutting-planes).
This approach shortens the solution process and outperforms standard techn
iques. Furthermore, we formally introduce a general concept of downgrading
that makes it possible to solve certain rostering problems that might other
wise have no solution. This gaper also computes a sample monthly schedule o
n the basis of a medium-sized European nil line's real data. (C) 2001 Elsev
ier Science Ltd. All rights reserved.