We can reformulate a vehicle routing problem (VRP) as an open shop scheduling problem (SSP) by representing visits as activities,
vehicles as resources on the factory floor, and travel as set up costs between activities. In this paper we present two reformulations:
from VRP to open shop, and the inverse, from SSP to VRP. Not surprisingly, VRP technology performs poorly on reformulated
SSP’s, as does scheduling technology on reformulated VRP’s. We present a pre-processing transformation that “compresses” the
VRP, transforming an element of travel into the duration of the visits. The compressed VRP’s are then reformulated as scheduling
problem, to determine if it is primarily distance in the VRP that causes scheduling technology to degrade on the reformulated
problem. This is a step towards understanding the features of a problem that make it more amenable to one technology rather
than another.
This work was supported by EPSRC research grant GR/M90641 and ILOG SA.