Shortest Path Problems in a Stochastic and Dynamic Environment

In this research, we consider stochastic and dynamic transportation network problems. Particularly, we develop a variety of algorithms to solve the expected shortest path problem in addition to techniques for computing the total travel time distribution along a path in the network. First, we develop an algorithm for solving an independent expected shortest path problem. Next, we incorporate the inherent dependencies along successive links in two distinct ways to find the expected shortest path. Since the dependent expected shortest path problem cannot be solved with traditional deterministic approaches, we develop a heuristic based on the K-shortest path algorithm for this dependent stochastic network problem. Additionally, transient and asymptotic versions of the problem are considered. An algorithm to compute a parametric total travel time distribution for the shortest path is presented along with stochastically shortest path measures. The work extends the current literature on such problems by considering interactions on adjacent links.