TY - GEN
T1 - Block-structured plan deordering
AU - Siddiqui, Fazlul Hasan
AU - Haslum, Patrik
PY - 2012
Y1 - 2012
N2 - Partially ordered plans have several useful properties, such as exhibiting the structure of the plan more clearly which facilitates post-plan generation tasks like scheduling the plan, explaining it to a user, or breaking it into subplans for distributed execution. The standard interpretation of partial ordering implies that whenever two subplans are unordered, every interleaving of steps from the two forms a valid execution. This restricts deordering to cases where individual steps (i.e., actions) are independent. We propose a weaker notion of partial ordering that divides the plan into blocks, such that the steps in a block may not be interleaved with steps outside the block, but unordered blocks can be executed in any sequence. We present an algorithm to find such deorderable blocks, and show that it enables deordering plans in many cases where no deordering is possible under the standard interpretation.
AB - Partially ordered plans have several useful properties, such as exhibiting the structure of the plan more clearly which facilitates post-plan generation tasks like scheduling the plan, explaining it to a user, or breaking it into subplans for distributed execution. The standard interpretation of partial ordering implies that whenever two subplans are unordered, every interleaving of steps from the two forms a valid execution. This restricts deordering to cases where individual steps (i.e., actions) are independent. We propose a weaker notion of partial ordering that divides the plan into blocks, such that the steps in a block may not be interleaved with steps outside the block, but unordered blocks can be executed in any sequence. We present an algorithm to find such deorderable blocks, and show that it enables deordering plans in many cases where no deordering is possible under the standard interpretation.
UR - http://www.scopus.com/inward/record.url?scp=84871377089&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-35101-3_68
DO - 10.1007/978-3-642-35101-3_68
M3 - Conference contribution
SN - 9783642351006
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 803
EP - 814
BT - AI 2012
T2 - 25th Australasian Joint Conference on Artificial Intelligence, AI 2012
Y2 - 4 December 2012 through 7 December 2012
ER -