ncm2 Distinguished Lecture
Grande conférence rcm2

Michel Balinski
CNRS & Laboratoire d'économétrie, École Polytechnique (Paris)

It is often said that equity is what is in the eyes of the beholder, but in
specific contexts persuasive principles of fairness may be invoked to obtain
practical solutions. The axiomatic approach proceeds by asking what properties
should be satisfied by any reasonable solution and then deducing the one method
(or the one class of methods) that satisfies them all ‹ unless too much is asked
and no method will do!

One such principle ‹ consistency ‹ generalizes the ideal of proportionality.

Its use is illustrated in several contexts:

o the problem of apportionment : how best to allocate the seats of a national
assembly among the provinces of a nation;

o the product variation problem : how best to sequence the production of items
in "just-in-time" production systems; and

o the problem of rounding : how best to round data.

Roughly speaking consistency says that any part of a solution should be a
solution : any part of a fair apportionment should be fair, any part of a good
production schedule should be good, any part of an accurate rounding should be
accurate ‹ just as corresponding sub-vectors of two proportional vectors are
proportional.

This may seem a minimal demand :  it leads, nonetheless, to very strong
conclusions.






Le jeudi 25 mai 2000 / Thursday, May 25, 2000
16 h /  4:00 p.m

HEC
3000, chemin de la Côte-Sainte-Catherine
Salle Luxembourg (1er étage sud, section jaune)
Réception après la conférence
(Salon Samson-Bélair, 4e étage, section bleue)






Centre de recherche en calcul appliqué (CERCA)
Centre interuniversitaire de recherche en analyse des organisations (CIRANO)
Centre de recherche informatique de Montréal (CRIM)
Centre de recherches mathématiques (CRM)
Centre de recherche sur les transports (CRT)
Groupe d¹études et de recherche en analyse des décisions (GERAD)
Institut national de la recherche scientifique - Télécommunications (INRS -
Télécommunications)