| Abstract: | There presently exists a gap between aggregate and disaggregate approaches to modeling spatial interaction behavior. This paper proposes a strategy for filling this gap by focusing on specific types of interaction behavior which share certain common characteristics. In the present case, a simple threshold theory is developed which establishes a link between various types of discretionary interaction behavior by individuals and the resulting macro distribution of interaction frequencies. In particular, it is shown that if the statistical population of potential interaction situations satisfies certain independence conditions over space, then the resulting macro interaction frequencies are always Poisson distributed with mean frequency levels representable by a classical gravity model. |
