In order to make planning technology more accessible and usable the planning community may have to adopt standard notations for embodying symbolic models of planning domains. In this paper it is argued that before we design such languages for planning we must be able to evaluate their quality. In other words, we must clear for what purpose the languages are to be used, and by what criteria the languages’ effectiveness are to be judged. Here some criteria are set down for languages used for theoretical and practical purposes respectively.
PDDL is evaluated with respect to them, with differing results depending on whether PDDL’s purpose is to be a theoretical or practical language. From the results of these evaluations some conclusions are drawn for the development
of standard languages for AI planning.
Downloads
Downloads per month over past year