08.07.2015 Views

Planning under Uncertainty in Dynamic Domains - Carnegie Mellon ...

Planning under Uncertainty in Dynamic Domains - Carnegie Mellon ...

Planning under Uncertainty in Dynamic Domains - Carnegie Mellon ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

List of Tables2.1 The policy iteration algorithm : : : : : : : : : : : : : : : : : : : : : : 152.2 The value iteration algorithm : : : : : : : : : : : : : : : : : : : : : : 163.1 Summary of the prodigy 4.0 plann<strong>in</strong>g algorithm. Each decisionpo<strong>in</strong>t is a potential backtrack<strong>in</strong>g po<strong>in</strong>t <strong>in</strong> the algorithm. : : : : : : : : 243.2 An annotated trace of prodigy 4.0 which shows a sequence of searchtree nodes created to produce the candidate plan shown <strong>in</strong> Figure 3.3.Each node except n5 is a child of the node <strong>in</strong> the l<strong>in</strong>e above. The text<strong>in</strong> italics has been added to the trace output for clarity of presentation. 263.3 The <strong>in</strong>itial state and goal description. Here t and f stand for \true"and \false". : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 384.1 An annotated trace of Prodigy, show<strong>in</strong>g the use of a protection node(n14) to add a precondition to the step at node n13 which will stop itsconditional eect from tak<strong>in</strong>g place. : : : : : : : : : : : : : : : : : : : 464.2 Algorithm for B-prodigy, based on prodigy 4.0. Steps which onlyappear <strong>in</strong> B-prodigy are shown <strong>in</strong> bold font. : : : : : : : : : : : : : 484.3 The aws uncovered by the plan critic for the rst candidate plan forthe example problem. For each aw a set of desired values is computedfrom the preconditions of the child action node, and the probabilitymass of this set is show. The explanations column shows parent nodevalues that can lead to an undesired value. : : : : : : : : : : : : : : 615.1 Algorithm to construct the belief net to evaluate a plan. : : : : : : : 766.1 Overview of the analogical replay procedure comb<strong>in</strong>ed with B-prodigy. 896.2 Operator schemas <strong>in</strong> the test doma<strong>in</strong>. : : : : : : : : : : : : : : : : : : 907.1 The objects and goal for the example problem used to demonstrate aconditional plan formed <strong>under</strong> scarce resources. : : : : : : : : : : : : 987.2 The <strong>in</strong>itial state for the example problem used to demonstrate a conditionalplan formed <strong>under</strong> scarce resources. : : : : : : : : : : : : : : 987.3 Prodigy's solution to the example problem, ignor<strong>in</strong>g <strong>in</strong>ference rules. : 99ix

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!