Risk-Aware Stochastic Shortest Path

We treat the problem of risk-aware control for stochastic shortest path (SSP) on Markov decision processes (MDP). Typically, expectation is considered SSP, which however oblivious to incurred risk. present an alternative view, instead optimizing conditional value-at-risk (CVaR), established risk measure. both chains as well MDP and introduce, through novel insights, two algorithms, based linear programming value iteration, respectively. Both algorithms offer precise provably correct solutions. Evaluation our prototype implementation shows that feasible several moderately sized models.