The 1998 AI Planning Systems Competition：（1998年的人工智能规划系统的竞争）.pdfVIP

The AI Planning Systems Comp etition Drew McDermott April Abstract The Planning Comp etition at the AI Planning Systems Confer ence was the rst of its kind Its goal was to create planning domains that a wide variety of planning researchers could agree on so as to make comparison among planners more meaningful measure overall progress in the eld and set up a framework for longterm creation of a rep ository of problems in a standard notation A rules committee for the comp etition was created in and had long discussions on how the contest should go One result of those discussions was the PDDL notation for planning domains This notation was used to set up a set of planning problems and to get a mo dest problem rep ository started As a result ve planning systems were able to comp ete when the contest to ok place in June All of these systems solved problems in the Strips framework with some slight extensions The attempt to nd domains for other forms of plan ning foundered b ecause of technical and organizational problems In spite of this the comp etition achieved its goals partially in that it conrmed that substantial progress had o ccurred in some subelds of planning and it allowed qualitative comparison among dierent planning algorithms It is urged that the comp etition continue to take place and to evolve History In recent years many subelds of AI have used comp etitions as a way of measur ing progress and guiding research directions eg In a comp etition researchers run their programs on a common set of problems at the same time with little tuning and the results are compared There are several purp oses to such an exercise It allows mea