Gerevini, Alfonso, Saetti, Alessandro and Vallati, Mauro (2012) PbP2: Portfolio-based Planner. University of Brescia, Brescia, Italy.
|Archive (ZIP) |
|Archive (ZIP) |
This is the code of the latest version of PbP, called PbP2.
It is a multi-planner approach that automatically
configures a portfolio of planning techniques for each given domain. The configuration process for a given domain uses a set of training instances to: (i) compute and analyze some alternative sets of macro-actions for each planner in the portfolio identifying a (possibly empty) useful set, (ii) select a cluster of planners, each one with an useful set of macro-actions, that is expected to perform best, and (iii) derive some additional information for configuring the execution scheduling of the selected planners at planning time.
PbP has two variants: PbP.s, focusing on speed, and PbP.q, focusing on plan quality.
A preliminary implementation of PbP entered the learning track of the sixth international planning competition (IPC6) and was the overall winner of this competition track. More recently, a revised and optimized version of PbP (PbP2) entered the learning track of the last competition (IPC7), and it was again the winner of this competition track.
|Subjects:||Q Science > QA Mathematics > QA76 Computer software|
|Schools:||School of Computing and Engineering|
School of Computing and Engineering > Informatics Research Group > Knowledge Engineering and Intelligent Interfaces
|Depositing User:||Mauro Vallati|
|Date Deposited:||13 Nov 2012 15:23|
|Last Modified:||06 Dec 2012 13:05|
Downloader CountriesMore statistics for this item...
Item control for Repository Staff only: