作者: David Chapman
DOI: 10.1016/0004-3702(87)90092-0
关键词: Operations research 、 Partial-order planning 、 Correctness 、 Heuristic 、 Simple (abstract algebra) 、 Action (philosophy) 、 Representation (mathematics) 、 Mathematics 、 Work (electrical) 、 State (computer science)
摘要: Abstract The problem of achieving conjunctive goals has been central to domain-independent planning research; the nonlinear constraint-posting approach most successful. Previous planners this type have complicated, heuristic, and ill-defined. I combined distilled state art into a simple, precise, implemented algorithm (TWEAK) which proved correct complete. analyze previous work on planning; in retrospect it becomes clear that all planners, linear nonlinear, same way. efficiency correctness these depends traditional add/delete-list representation for actions, drastically limits their usefulness. present theorems suggest efficient general purpose with more expressive action representations is impossible, ways avoid problem.