Fuzzy Vs. Random Utility Models: an Application to Mode Choice Behaviour Analysis
G E Cantarella, University of Salerno; V Fedele, Mediterranean University of Reggio Calabria, IT
The most common behavioural paradigm underlining discrete choice analysis assumes that:
A. the decision-maker, e.g. each user of a transportation system facing the choice of the transportation mode,
a.1 considers a set of mutually exclusive alternatives,
a.2 gives each alternative a value of perceived utility,
a.3 chooses the (an) alternative with the maximum value of perceived utility;
B. the perceived utility of each alternative is modelled taking into account sources of uncertainty, regarding information available to the user as well as to the modeller. The utility is expressed as a function of attributes, which may be measured or assumed in a design scenario, and parameters, which are estimated through a sample of observations.
If the perceived utility is modelled through a random variable the well-known Random Utility Theory is obtained. Models derived from such a theory have been applied with satisfactory results over the years.
This paper discusses the Fuzzy Utility Theory resulting from assuming that the perceived utility is modelled through a fuzzy number. An application to mode choice analysis is also presented for a real case, showing that Fuzzy Utility Models are feasible
Association for European Transport