This paper proposes a dynamic programming (DP) approach for a stochastic multi-period allocation problem, whereby fleet of vehicles are assigned from stations to destinations with multiple alternative routes in order to maximize returns, while the vehicles are subject to random failure. In the process of managing the business, the company is assumed to incur proportional management costs and pay tax to government. The expected returns is modelled as a function of random failure of vehicles due to bad roads and depreciation. The depreciation rate is assumed to follow a straight-line approach. The breakdown rate is modelled as function of the rate of bad roads on the fleets, depreciation rate... More >