Trace:

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision 		Previous revision
Last revision Both sides next revision
general_purpose_planners_survey [2010/10/15 09:08]
jakub_gemrot 		general_purpose_planners_survey [2011/04/20 15:57]
jakub.gemrot
Line 18: Line 18:
 <td style="border-bottom: 1px solid lightgray;">Excitement:</td> <td style="border-bottom: 1px solid lightgray;">Excitement:</td>
 <td> </td> <td> </td>
-<td><font color="blue">There already exist a vast number of general purpose planners, but hey, it has been said that these planners are optimized to only certain competition-problems and compete differently on equivalent problems that are just reformulated (even changing names of operators may have disastrous effects!). Wow, we should check that and tell the world that all these planners are strange and weird and their measured performance is totally different from what everyone knows!</font></td>+<td><font color="blue">Tell the world how the planners are truly behaves!</font></td>
 </tr> </tr>
 <tr> <tr>
 <td style="border-bottom: 1px solid lightgray;">Description:</td> <td style="border-bottom: 1px solid lightgray;">Description:</td>
 <td> </td> <td> </td>
-<td><font color="default">+<td><font color="black">
 You will work with 5-10 general purpose planners that you should test against well defined problems described in PDDL. The main work will lie in the reformulation of various problems and creating a program that will automatically generate an equivalent problems so you may effectively test performace of planners on bunch of different (yet the same) problems to gain insight in their true performance. Note that the program won't be hard to create - the aim is not to devise whole academic framework on "how to generate equivalent problems", but to generate problems that have for instance different operator names of arity, etc. The analysis of results will conclude your work reporting on strengths and weaknesses of various planners. You will work with 5-10 general purpose planners that you should test against well defined problems described in PDDL. The main work will lie in the reformulation of various problems and creating a program that will automatically generate an equivalent problems so you may effectively test performace of planners on bunch of different (yet the same) problems to gain insight in their true performance. Note that the program won't be hard to create - the aim is not to devise whole academic framework on "how to generate equivalent problems", but to generate problems that have for instance different operator names of arity, etc. The analysis of results will conclude your work reporting on strengths and weaknesses of various planners.
 <br/><br/> <br/><br/>
general_purpose_planners_survey.txt · Last modified: 2012/02/11 18:16 by jakub.gemrot