Vehicle sizing model for bus transit networks

An iterative procedure to select vehicle sizes for the routes of a bus system with a given network configuration and origin-destination demand matrix is presented. The procedure starts by assigning a set of initial route service frequencies to compute route-level descriptors through a transit trip assignment model. The vehicle size for each route is computed analytically by a mathematical model that minimizes the total cost (operator cost and user cost) of each individual bus route. Revised frequencies are determined by applying a maximum allowed load factor consistent with the calculated vehicle sizes. The procedure terminates when frequencies of two consecutive iterations converge. The model is illustrated through a case application to the existing transit system in Austin, Texas. The result confirms the potential benefits of using variable vehicle sizes on different routes. However, the number of vehicle sizes in a system should be limited to avoid operational complexity and associated maintenance costs. In general, it appears that smaller buses could be operated on most of the bus routes in most North American cities to provide better service quality and lower operator cost.