Optimal paths in graphs with stochastic or multidimensional weights

Abstract

This paper explores computationally tractable formulations of stochastic and multidimensional optimal path problems, each as an extension of the shortest path problem. A single formulation encompassing both problems is considered, in which a utility function defines preference among candidate paths. The result is the ability to state explicit conditions for exact solutions using standard methods, and the applicability of well-understood approximation techniques.