Heuristic Search for Multi-Objective Probabilistic Planning

Heuristic search is a powerful approach that has successfully been applied to broad class of planning problems, including classical planning, multi-objective and probabilistic modelled as stochastic shortest path (SSP) problem. Here, we extend the reach heuristic more expressive namely paths (MOSSPs), which require computing coverage set non-dominated policies. We design new algorithms MOLAO* MOLRTDP, well-known SSP case. further construct spectrum domain-independent functions differing in their ability take into account features problem guide search. Our experiments demonstrate benefits these relative merits heuristics.