Determination of the Uncertainty Derived from Input Data in Path Finding Analysis Output

Management of some facilities such as transportation is one of the most important applications of geographic information systems. In this field, the network capabilities of GIS such as computation of shortest path could be very useful. There are a variety of studies and researches on developing algorithms for shortest path finding. However, less attention has been paid on modeling the uncertainty of determined optimum path. The uncertainty of a determined optimum path is relevant to the accuracy of the input data. For example, when distance is considered as a criterion for determination of optimum path, the positional accuracy of the road network affect the reliability of the determined optimum path. This paper intends to model the uncertainty of optimum paths, where the optimality is determined based on distance and travel time criteria. Error propagation rules and fuzzy set theory were used respectively to model the uncertainty for distance criteria and travel time criteria. The developed model was implemented and tested for a case study in Iran. The results showed that by including uncertainty in a path finding problem, users can determine optimum paths, which may be a little bit longer (from length or travel time perspectives), but more reliable than the shortest or the fastest path. Using a reliable optimum path is preferred by many citizens (and especially emergency services) in comparison with uncertain shortest/fastest path.