Mapping contingent planning into multi-valued planning

Contingent planning involves generating plans in uncertainty environment under partial observability. However, traditional contingent planning problems, described by propositional logic, suffer from many problems. For example, the belief state space is too large. In this paper, we introduce a novel approach that can map propositional contingent planning tasks into multi-valued contingent planning task (MCPT). Experimental results have shown that our approach is sound and efficient.