A GIS-based stochastic approach to generating daytime population distributions for vehicle route planning

Abstract

This article explores the use of daytime population distributions for vehicle route planning in urban environments when proximity to population is a major consideration. A GIS-based stochastic geocoding algorithm is proposed to generate daytime population distribution patterns using travel survey data. Metropolitan Melbourne in Australia is used as a test case with data collected from the Victorian Activity and Travel Survey between 1994 and 1999. The advantage of employing daytime population distributions for vehicle route planning is then illustrated taking dangerous goods delivery and mobile billboard advertising as examples. Analysis suggests that the proposed algorithm generates population distribution patterns with finer spatial and temporal details than that obtained using other estimation methods. The algorithm also eliminates the problem of over- or under-estimation of population associated with previous attempts using travel survey data and therefore provides more accurate results. © 2009 Blackwell Publishing Ltd.