Reduction of greenhouse gas (GHG) emissions from ground transportation is increasingly becoming one of the most important issues for automobile manufacturers to achieve environmental sustainability in the transportation sector. As a result, recent research has paid considerable attention to the use of alternative-fuel vehicles (AFVs) that may help reduce GHG emissions. However, due to the lack of an initial infrastructure, which would include alternative-fuel refueling stations, the use of AFVs is still in their early stages. In order to commercialize AFVs in a given region, it is necessary to set up an initial set of refueling stations on the region's road network. In this paper, we develop a mathematical model to determine optimal locations for a pre-determined number of refueling stations whose objective is to maximize the traffic flows covered on a tree network, which is a common structure of toll roads in many states. After collecting the necessary origin/destination traffic flow data, the model is applied to the Pennsylvania Turnpike network.