Residential location choice modelling : gaussian distributed stochastic utility functions

Synopsis

Modelling urban residential location choice using the theory of constrained utility maximising behaviour by individual locators requires a probabilistic approach.   One such approach using random utility theory assumes utilities are stochastic but that decision making is deterministic.   This approach can only be made operational by adopting explicit distributional forme for the stochastic utility functions. The extreme value type I (Gumbel) distribution is almost universally adopted, principally for mathematical convenience rather than for any particular conceptual reasons.   This paper examinee the derivation of a choice model based on Gaussian distributed stochastic utility functions. Expressions for locator's realised utility are also developed. The presentation is confined to the case of mutually independent alternatives.  Although the paper specifically addressee urban residential location choice, the results are equally applicable in the context of retail shopping models or trip distribution models.