Abstract

An analysis of strategies, recognizable abstract patterns of planned behavior, reveals an enormous divide between the kind of planning that artificial intelligence planning systems do and the kind of strategic planning that people do. This paper describes a project to collect and represent strategies on a large scale to identify the representational requirements of strategic planning in order to inform the design of future artificial intelligence planning systems. Three hundred and seventy-two strategies were collected from ten different planning domains. Each was represented at an abstract level and in a preformal manner designed to reveal the planning concepts that each strategy contains. The contents of these representations, consisting of nearly one thousand unique planning concepts, were then collected and organized into forty-eight groups that outline the representational and functional components of strategic planning systems.

Citations