Accessibility is a very important influence in urban space, and it similarly plays an important role in UrbanSim. Almost all models in UrbanSim consider the effects of accessibility. But unlike the monocentric or spatial interaction models, in which the choice of workplace is exogenous and residential locations are chosen principally on the basis of commute to the city center or to a predetermined workplace, we deal with accessibility in a more general framework. Accessibility is considered a normal good, like other positive attributes of housing, which consumers place a positive economic value on. We therefore expect that consumers value access to workplaces and shopping opportunities, among the many other attributes they consider in their housing preferences. However, not all households respond to accessibility in the same way. Retired persons would be less influenced by accessibility to job opportunities than would working age households, for instance.
We operationalize the concept of accessibility for a given location as the distribution of opportunities weighted by the travel impedance, or alternatively the utility of travel to those destinations. A number of alternative accessibility measures have been developed in UrbanSim. The utility of travel is measured as the composite utility across all modes of travel for each zone pair, obtained as the logsum of the mode choice for each origin-destination pair.
The accessibility model reads the logsum matrix from the travel model and the land use distribution for a given year, and creates accessibility indices for use in the household and business location choice models. The general framework is to summarize the accessibility from each zone to various activities for which accessibility is considered important in household or business location choice.
Since UrbanSim operates annually, but travel model updates are likely to be executed for two to three of the years within the forecasting horizon, travel utilities remain constant from one travel model run until they are replaced by the next travel model result. Although travel utilities remain constant, the activity distribution in these accessibility indices is updated annually, so that the accessibility indices change from one year to the next to reflect the evolving spatial distribution of activities.