Solution of an RMDP
	• Solve a complete domain in one go. • Output a mapping from an abstract state (set of states) to a policy. • Alternatively, output a function model to map a state to a value (like say linear combination of objects)