Permutation Problems and Channelling Constraints

Many  assignment,  scheduling  and routing  problems  are  permutation
problems.  For example, sports tournament scheduling can be modeled as
finding a permutation  of the games to fit into the  available  slots,
whilst the traveling  salesperson  problem can be modeled as finding a
permutation of the cities which gives a tour of minimal length.  There
are a number of different ways of modelling  permutation problems (for
example, do we assign  teams to time slots, or do we assign time slots
to  teams?).  I  describe  a  methodology,  as well  as a  measure  of
constraint  tightness,  that can be used to  compare  these  different
models.  My  results  will  help  users  of  constraints  satisfaction
toolkits  to choose a model  for a  permutation  problem,  and a local
consistency.