A note on exact algorithms for the bottleneck generalized assignment problem