Computational models of planning

Abstract

The selection of the action to do next is one of the central problems faced by autonomous agents. Natural and artificial systems address this problem in various ways: action responses can be hardwired, they can be learned, or they can be computed from a model of the situation, the actions, and the goals. Planning is the model-based approach to action selection and a fundamental ingredient of intelligent behavior in both humans and machines. Planning, however, is computationally hard as the consideration of all possible courses of action is not computationally feasible. The problem has been addressed by research in Artificial Intelligence that in recent years has uncovered simple but powerful computational principles that make planning feasible. The principles take the form of domain-independent methods for computing heuristics or appraisals that enable the effective generation of goal-directed behavior even over huge spaces. In this paper, we look at several planning models, at methods that have been shown to scale up to large problems, and at what these methods may suggest about the human mind. WIREs Cogn Sci 2013, 4:341-356. doi: 10.1002/wcs.1233 The authors have declared no conflicts of interest for this article. For further resources related to this article, please visit the WIREs website.

The execution of planned detours by spider-eating predators

“…More specifically, we were interested in whether test spiders, after leaving the Tower, used representations that they had derived while on top of the Tower. Another way of saying this is that we were interested in whether our findings have at least a rudimentary correspondence to Geffner's () thinking‐before‐acting characterization of planning.…”

“…Nonetheless, spiders sometimes behave in a way that makes at least a rudimentary capacity for planning difficult to deny. Geffner (2013) proposed that there are three basic ways by which an individual might make plans, with these ways being aligned with Dennett's (1995Dennett's ( , 1996 distinctions between Darwinian, Skinnerian, and Popperian animals (also see Raby & Clayton, 2009). A Darwinian animal relies on what Geffner called a "hardwired approach", with the animal's "innate" or "instinctive" (Lorenz, 1965) plan being derived by natural selection, a trial-and-error process acting over evolutionary time (e.g., see Catania, 2010).…”

Automated Planning for Business Process Management

“…There exist several forms of planning models in the AI literature, which result from the application of some orthogonal dimensions [34]: uncertainty in the initial state (fully or partially known) and in the actions dynamics (deterministic or not), the type of feedback (full, partial or no state feedback) and whether uncertainty is represented by sets of states or probability distributions.…”

Artificial Intelligence: From programs to solvers

“…I thank Michael Fourman, Johanna Moore, Andy Barto and Shlomo Zilberstein for the invitations and their hospitality. I have touched on some of these issues in [21,23]. My work is partially supported by grant EC-7PM-SpaceBook.…”