We describe the ucpop partial order planning algorithm which handles a subset of Pednault's ADL action representation. In particular, ucpop operates with actions that have conditional effects, universally quantified preconditions and effects, and with universally quantified goals. We prove ucpop is both sound and complete for this representation and describe a practical implementation that succeeds on all of Pednault's and McDermott's examples, including the infamous "Yale Stacking Problem" [McDermott 1991].

We develop a formal model of planning based on satisfiability rather than deduction. The satis ability approach not only provides a more flexible framework for stating di erent kinds of constraints on plans, but also more accurately reflects the theory behind modern constraint-based planning systems. Finally, we consider the computational characteristics of the resulting formulas, by solving them with two very different satisfiability testing procedures.

The research on conditional planning rejects the assumptions that there is no uncertainty or incompleteness of knowledge with respect to the state and changes of the system the plans operate on. Without these assumptions the sequences of operations that achieve the goals depend on the initial state and the outcomes of nondeterministic changes in the system. This setting raises the questions of how to represent the plans and how to perform plan search. The answers are quite different from those in the simpler classical framework. In this paper, we approach conditional planning from a new viewpoint that is motivated by the use of satisfiability algorithms in classical planning. Translating conditional planning to formulae in the propositional logic is not feasible because of inherent computational limitations. Instead, we translate conditional planning to quantified Boolean formulae. We discuss three formalizations of conditional planning as quantified Boolean formulae, and pr...

Intelligent agents are systems that have a complex, ongoing interaction with an environment that is dynamic and imperfectly predictable. Agents are typically difficult to program because the correctness of a program depends on the details of how the agent is situated in its environment. In this paper, we present a methodology for the design of situated agents that is based on situated automata theory. This approach allows designers to describe the informational content of an agent's computational states in a semantically rigorous way without requiring a commitment to conventional runtime symbolic processing. We start by outlining this situated view of representation, then show how it contributes to design methodologies for building systems that track perceptual conditions and take purposeful actions in their environments. 1 Introduction Humans, delivery robots, and automated factories are all systems that have an intelligent, ongoing interaction with environments that are dynamic and ...

We are concerned with the implications and interactions of three common expressive extensions to classical planning: conditional plans, context-dependent actions, and nondeterministic action outcomes. All of these extensions have appeared in recent work, sometimes in conjunction, but the semantics of the combination has not been fully explored. As we have argued in previous work, providing a coherent semantics for conditional planning with context-dependent actions requires that the planner's information state be modelled separately from the world state. In this paper, we present a new planning language, WCPL, encompassing these extensions. The semantics of WCPL includes an explicit treatment of the planner's information state as knowledge, as opposed to some form of context labelling. In addition to clarifying and unifying a disparate set of results from earlier work, we extend that work: WCPL handles both conditional and fail-safe plans for an action representation including both context-dependent and nondeterministic actions.

We study an automata-theoretic approach to planning for temporally extended goals. Specifically, we devise techniques based on nonemptiness of Büchi automata on infinite words, to synthesize sequential and conditional plans in a generalized setting in which we have that: goals are general temporal properties of desired execution; dynamic systems are represented by finite transition systems; incomplete information on the initial situation is allowed; and states are only partially observable. We prove that the techniques proposed are optimal wrt the worst case complexity of the problem. Thanks to the scalability of the nonemptiness algorithms, the techniques presented here promise to be applicable to fairly large systems, notwithstanding the intrinsic complexity of the problem.

One way to think about STRIPS is as a mapping from databases to databases, in the following sense: Suppose we want to know what the world would be like if an action, represented by the STRIPS operator ff, were done in some world, represented by the STRIPS database D 0 . To find out, simply perform the operator ff on D 0 (by applying ff's elementary add and delete revision operators to D 0 ). We describe this process as progressing the database D 0 in response to the action ff. In this paper, we consider the general problem of progressing an initial database in response to a given sequence of actions. We appeal to the situation calculus and an axiomatization of actions which addresses the frame problem (Reiter [13], Lin and Reiter [8]). This setting is considerably more general than STRIPS. Our results concerning progression are mixed. The (surprising) bad news is that, in general, to characterize a progressed database we must appeal to second order logic. The good news is that there...

This article gives an overview of AI (Artificial Intelligence) plan-ning techniques and discusses their application to the Web service composition problem.

The focus of current research in cognitive robotics is both on the realization of systems based on known formal settings and on the extension of previous formal approaches to account for features that play a significant role for autonomous robots, but have not yet received an adequate treatment. In this paper we adopt a formal framework derived from Propositional Dynamic Logics by exploiting their formal correspondence with Description Logics, and present an extension of such a framework obtained by introducing both concurrency on primitive actions and autoepistemic operators for explicitly representing the robot’s epistemic state. We show that the resulting formal setting allows for the representation of actions with contextdependent effects, sensing actions, and concurrent actions, and address both the presence of exogenous events and the characterization of the notion of executable plan in such a complex setting. Moreover, we present an implementation of this framework in a system which is capable of generating plans that are actually executed on mobile robots, and illustrate the experimentation of such a system in the design and implementation of soccer players for the 1999 Robocup competition.

In this paper we present a logical framework for defining consistent axiomatizations of planning domains. A language to define basic actions and structured plans is embedded in a logic. This allows general properties of a whole planning scenario to be proved as well as plans to be formed deductively. In particular, frame assertions and domain constraints as invariants of the basic actions can be formulated and proved. Even for complex plans most frame assertions are obtained by purely syntactic analysis. In such cases the formal proof can be generated in a uniform way. The formalism we introduce is especially useful when treating recursive plans. A tactical theorem prover, the Karlsruhe Interactive Verifier KIV is used to implement this logical framework. 1 Introduction In this paper we present a logical framework for defining consistent axiomatizations of planning domains. An effective mechanism for defining the basic operations in a constructive way is embedded in a logic that allo...