If A is an n x n nonnegative, irreducible matrix, then there exists mu
(A) > 0, and a positive vector x such that max(j)a(ij)x(j) = mu(A)x(i)
, i = 1, 2, ..., n. Furthermore, mu(A) is the maximum geometric mean o
f a circuit in the weighted directed graph corresponding to A. This th
eorem, which we refer to as the max version of the Perron-Frobenius Th
eorem, is well-known in the context of matrices over the max algebra a
nd also in the context of matrix scalings. In the present work, which
is partly expository, we bring out the intimate connection between thi
s result and the Perron-Frobenius theory. We present several proofs of
the result, some of which use the Perron-Frobenius Theorem. Structure
of max eigenvalues and max eigenvectors is described. Possible ways t
o unify the Perron-Frobenius Theorem and its max version are indicated
. Some inequalities for mu(A) are proved. (C) 1998 Elsevier Science In
c. All rights reserved.