STATIONARY OPTIMAL POLICIES WITH DISCOUNTING IN A STOCHASTIC ACTIVITY ANALYSIS MODEL