P. Glasserman et Dd. Yao, STOCHASTIC VECTOR DIFFERENCE-EQUATIONS WITH STATIONARY COEFFICIENTS, Journal of Applied Probability, 32(4), 1995, pp. 851-866
We give a unified presentation of stability results for stochastic vec
tor difference equations Y-n+1 = A(n)X Y-n+B-n based on various choice
s of binary operations + and X, assuming that {(A(n), B-n), n greater
than or equal to O} are stationary and ergodic. In the scalar case, un
der standard addition and multiplication, the key condition for stabil
ity is E[log\A(0)\]<O. In the generalizations, the condition takes the
form gamma<0, where gamma is the limit of a subadditive process assoc
iated with {A(n), n greater than or equal to O}. Under this and mild a
dditional conditions, the process {Y-n, n greater than or equal to O}
has a unique finite stationary distribution to which it converges from
all initial conditions. The variants of standard matrix algebra we co
nsider replace the operations + and x with (max,+), (max, x), (min,+),
or (min, x). In each case, the appropriate stability condition parall
els that for the standard recursions, involving certain subadditive li
mits. Since these limits are difficult to evaluate, we provide bounds,
thus giving alternative, computable conditions for stability.