(2003), Crainic and Kim (2007), Günther and Kim (2005) and Steenken et al. (2004)). Most are based on static formulations
using the carrier’s historic data and forecasts. The advent of ITS location and communication technologies offers the possibility
to dramatically enhance the quantity and quality of the data available for the forecast and planning processes. This
should translate into better plans and operations that are more profitable.
Parallel and distributed computing is an enabling factor for ITS in general and CVO-AFMS in particular. Its challenges are
of two different but complementary natures. On the one hand, parallel computing offers the possibility to design data analysis
and decision-support system architectures to answer efficiently complex requests in real or quasi-real time. Thus, processors
may be dedicated to the various tasks of receiving, validating, and formatting data, analysing and aggregating it,
forecasting, background simulation-optimisation, real-time selection of the appropriate strategy, etc. On the other hand, parallel
computing also offers a challenging perspective with potentially great rewards: to solve realistically formulated and
dimensioned problem instances within reasonable times. Each class of problems and algorithms presents its own challenges.
It appears clearly, however, that research efforts have to be dedicated both to the decomposition and distribution of tasks
corresponding to one particular problem instance and algorithm, and to the development of co-operating search mechanisms
that bring to bear on any given problem instance the combined power of several exact methods and meta-heuristics.