Suppose that X-1, X-2,... is a stationary stochastic process of positive k
x k matrices, and let Y-n(1) = (XXn-1)-X-n ...X-1 be the corresponding prod
uct matrices. For a special case, Bellman showed that the elements [Y-n(1)]
(ij) converge in the sense that n(-1)E{log[Y-n(1)](ij)} --> a as n --> infi
nity. The constant a is independent of i and j. Bellman also conjectured th
at, asymptotically, the n(-1/2){log[Y-n(1)](ij) - na} terms are distributed
according to a normal distribution with a common variance, independent of
ij. Later Furstenberg and Kesten generalized and strengthened Bellman's res
ult and established the validity of his conjecture.
This paper extends these results to the case of nonlinear mappings that are
monotonic and homogeneous of degree one on R-+(k). Specifically, given a s
tationary process H-1, H-2,... of such mappings, we define the composite ma
ppings F-n(1)(.) = H-n (Hn-1(...(H-1(.)...). Under appropriate conditions,
the components [F-n(1)(x(0))](i) have the property that, almost surely, n(-
1) log[F-n(1)(x(0))](i) --> a independent of x(0) and i. Furthermore the co
mponents n(-1/2){log[F-n(1)(x(0))](i) - na} are asymptotically distributed
according to a normal distribution with a common variance.