Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 52027 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): GhassemSani, Gholamreza
- Abstract:
- Successful performance of forward heuristic planners in classical planning, has motivated the use of forward heuristic search in many temporal planners. Most of forward temporal planners are decision epoch planners that use an explicit representation of time and schedule actions while searching. However, decision epoch planners aren’t complete and cannot solve certain temporal problems with required concurrency. In this dissertation, we present ATFP, a forward temporal planner that does not use an explicit representation of time. In ATFP, duration of actions are temporarily ignored, and a causally valid order of beginnings and endings of actions is searched using classical forward heuristic search. Only after finding a causally valid sequence of actions, the necessary temporal conditions of a valid temporal plan are checked. If the obtained causally valid plan is also temporally consistent, it will be scheduled to form the output plan. Otherwise, the cause of temporal inconsistency will be detected, so that it can be used to prune the search. Experimental results show that ATFP’s performance is comparable to temporal planners like ITSAT and POPF that are capable of solving temporal problems with required concurrency
- Keywords:
- Temporal Planning ; Forward State Space Planning ; Required Concurrency ; Time Abstraction
Digital Object List
-
محتواي کتاب
- view
Bookmark