机构地区: 暨南大学信息科学技术学院计算机科学系
出 处: 《计算机科学》 2007年第3期176-180,共5页
摘 要: 派生谓词是PDDL2.2语言的新特性之一。在2004年的规划大赛IPC-4上,许多规划系统都无法求解包含派生谓词的两个标准竞赛问题。在经典规划中,派生谓词是指不受领域动作直接影响的谓词,它们在当前状态下的真值是在封闭世界假设中由某些基本谓词通过领域公理推导出来的。本文提出一种新的方法来求解包含派生谓词的规划问题,即用与状态无关的激活集来取代派生谓词用于放宽式规划中。 Derived predicate is one of two new features of PDDL2.2 language, and two domains with derived predicates in IPC-4 are very challenging to many planning system. In classical planning, derived predicates are predicates that are not effected directly by the domain actions, and their truth in the current state is inferred from that of some basic predicates via some domain axioms under the closed world assumption. Similar to rule graph and state-dependent activation set proposed by LPG-td, we propose another approach to planning with derived predicates where state-independent (not state-dependent)activation sets of a derived predicate are used in relax-plan heuristics.
关 键 词: 智能规划 派生谓词 规则图 激活集 放宽式规划
领 域: [自动化与计算机技术] [自动化与计算机技术]