Vehicle scheduling can be a simple paper exercise or can become so complex that computer based mathematical models have to be used. Having a transport supervisor or manager plan the delivery schedule on the basis of a detailed knowledge of each driver's capabilities, the geographical area and the particulars of the customers have inherent advantages. However, fleets quickly become too large for individual personnel to produce scheduling plans which maximise efficiency and effectiveness at minimum cost. Moreover, in highly competitive markets, physical distribution costs have to be carefully managed if the marketing margin is not to be entirely eroded. Thus, to some point, most organisations will have to give consideration to applying mathematical models to vehicle scheduling decisions. In most instances, these mathematical models will prove unready unless they are manipulated by a computer.