Shortest Path Based Decision Making Using Probabilistic Inference

Recently, there is an increasing interest in solving sequential The shortest path routing (SPR) problem entails finding the decision making problems under uncertainty, such shortest path between two given vertices of the graph that as Markov decision processes (MDPs), partially observable minimizes the total sum of weights of the involved edges MDPs (POMDPs) and its multiagent variants using in the path (Dijkstra 1959; Cormen et al. 2001). Shortest the probabilistic inference based viewpoint (Toussaint and path based approaches have found several applications in Storkey 2006; Toussaint, Charlin, and Poupart 2008; Kumar diverse fields such as transportation models (Pallottino and and Zilberstein 2010; Ghosh, Kumar, and Varakantham Scutella 1998), telecommunication network design (Pioro et 2015; Kumar, Zilberstein, and Toussaint 2015). Our al. 2002) and in ecology for analyzing landscape connectivity approach for SPDM is different from such previous applications for conservation planning (Bunn, Urban, and Keitt 2012; of the LM framework for planning.