One of the ways to solve the classical apportionment problem (which has been studied chiefly in relation to the apportionment of seats in a chamber of representatives) is the optimization of a discrepancy function; although this approach seems very natural, it has been hardly used. In this paper, we propose a more general formalization of the problem and an optimization procedure for a broad class of discrepancy functions, study the properties of the procedure and present some examples in which it is applied.

Peer Reviewed