No label defined (Q4704802)

The paper deals with what the authors call share functions, that are, essentially, artificial normalizations of values intended to make them efficient in some sense (as the payoffs to the players sum to 1 for any share function). The domain is restricted to (nonnull, when necessary) monotonic games, although it might be easily enlarged at some extent. Of course, the most conspicuous example to which the idea applies is the Banzhaf value. (Incidentally, the authors give erroneously credit to \textit{J. F. Banzhaf} [Weighted voting doesn't work: A mathematical analysis. Rutgers Law Review 19, 317-343 (1965)], instead of \textit{G. Owen} [Nav. Res. Logist. Q. 22, 741-750 (1975; Zbl 0333.90056)], for the definition of the Banzhaf value on the space of all cooperative games.) A series of definitions, axioms and statements (most of them giving axiomatic characterizations) is offered along the paper. All of them are systematically preceded or followed by intuitive comments that make the paper highly readable. In fact, continuous references to three previous papers by the authors are given [\textit{R. van den Brink} and \textit{G. van der Laan}, Axiomatization of the normalized Banzhaf value and the Shapley value. TI-Discussion Paper 8-95-194, Tilburg University (1995) to appear in Social Choice Welfare); Axiomatizations of a class of share functions for \(n\)-person games. TI-Discussion Paper 8-95-154, Tilburg University; see also Theory Decis. 44, 117-148 (1998; Zbl 0905.90190); Potentials, share functions and core concepts for share vectors, Mimeo (1998)], to the extent that the text becomes somewhat reiterative and it is not at all clear which is the strictly new contribution of the paper. Therefore the readers will probably split into two groups: those that, being aware of these three previous papers, will find almost no new idea on the topic, and those that, ignoring the authors' previous work, will appreciate the paper as an interesting compilation of results.