AI planning agents are goal-directed: success is measured in terms of whether or not an input goal is satisfied, and the agent's computational processes are driven by those goals. A decision-theoretic agent, on the other hand, has no explicit goals--- success is measured in terms of its preferences or a utility function that respects those preferences. The two approaches have complementary strengths and weaknesses. Symbolic planning provides a computational theory of plan generation, but under unrealistic assumptions: perfect information about and control over the world and a restrictive model of actions and goals. Decision theory provides a normative model of choice under uncertainty, but offers no guidance as to how the planning options are to be generated. This paper unifies the two approaches to planning by describing utility models that support rational decision making while retaining the goal information needed to support plan generation. We develop an extended model of goals tha...

One big obstacle to understanding the nature of hierarchical task network (htn) planning has been the lack of a clear theoretical framework. In particular, no one has yet presented a clear and concise htn algorithm that is sound and complete. In this paper, we present a formal syntax and semantics for htn planning. Based on this syntax and semantics, we are able to de ne an algorithm for htn planning and prove it sound and complete. We also develop several de nitions of expressivity for planning languages and prove that htn planning is strictly more expressive than strips-style planning according to those de nitions. 1

Many potential uses of qualitative physics, such as robot planning and intelligent computer-aided engineering, require integrating physics with actions taken by agents. This paper proposes to augment qualitative simulation to include the effects of actions to form action-augmented envisionments. The action-augmented envisionment incorporates both the effects of an agent's actions and what will happen in the physical world whether or not the agent does something. Consequently, it should provide a richer basis for planning and procedure generation than any previous representation. This paper defines actionaugmented envisionments and an algorithm for directly computing them, along with an analysis of its complexity and suitability for different kinds of problems. We describe our initial implementation and discuss potential extensions, including incremental algorithms. Keywords: Qualitative reasoning, planning, artificial intelligence. Presented at the 2nd Qualitative Physics Workshop Pa...

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.

This article studies the problem of modifying the action ordering of a plan in order to optimise the plan according to various criteria. One of these criteria is to make a plan less constrained and the other is to minimize its parallel execution time. Three candidate definitions are proposed for the first of these criteria, constituting a sequence of increasing optimality guarantees. Two of these are based on deordering plans, which means that ordering relations may only be removed, not added, while the third one uses reordering, where arbitrary modifications to the ordering are allowed. It is shown that only the weakest one of the three criteria is tractable to achieve, the other two being NP-hard and even difficult to approximate. Similarly, optimising the parallel execution time of a plan is studied both for deordering and reordering of plans. In the general case, both of these computations are NP-hard. However, it is shown that optimal deorderings can be computed in polynomial time...

The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the o cial policies, either expressed or implied, of any other parties.

Artificial intelligence research falls roughly into two categories: formal and implementational. This division is not completely firm: there are implementational studies based on (formal or informal) theories (e.g., CYC, SOAR, OSCAR), and there are theories framed with an eye toward implementability (e.g., predicate circumscription). Nevertheless, formal /theoretical work tends to focus on very narrow problems (and even on very special cases of very narrow problems) while trying to get them "right" in a very strict sense, while implementational work tends to aim at fairly broad ranges of behavior but often at the expense of any kind of overall conceptually unifying framework that informs understanding. It is sometimes urged that this gap is intrinsic to the topic: intelligence is not a unitary thing for which there will be a unifying theory, but rather a "society" of subintelligences whose overall behavior cannot be reduced to useful characterizing and predictive principles.

Abstract. In this paper we describe Constraint-based Attribute and Interval Planning (CAIP), a paradigm for representing and reasoning about plans. The paradigm enables the description of planning domains with time, resources, concurrent activities, mutual exclusions among sets of activities, disjunctive preconditions and conditional effects. We provide a theoretical foundation for the paradigm, based on temporal intervals and attributes. We then show how the plans are naturally expressed by networks of constraints, and show that the process of planning maps directly to dynamic constraint reasoning. In addition, we define compatibilities, a compact mechanism for describing planning domains. We describe how this framework can incorporate the use of constraint reasoning technology to improve planning. Finally, we describe EUROPA, an implementation of the CAIP framework. 1. What Should a Planner Do? In recent years, planning has been applied to complex domains, including the sequencing of commands for spacecraft both on the ground and on-board (Jónsson et al., 2000). The domain of spacecraft operations

Many real-world planning problems require generating plans that maximize the parallelism inherentina problem. There are a number of partial-order planners that generate such plans# however, in most of these planners it is unclear under what conditions the resulting plans will be correct and whether the planner can even find a plan if one exists. This paper identifies the underlying assumptions about when a partial plan can be executed in parallel, defines the classes of parallel plans that can be generated by different partial-order planners, and describes the changes required to turn ucpop into a parallel execution planner. In addition, we describe how this planner can be applied to the problem of query access planning, where parallel execution produces substantial reductions in overall execution time.

VHPOP is a partial order causal link (POCL) planner loosely based on UCPOP. It draws from the experience gained in the early to mid 1990’s on flaw selection strategies for POCL planning, and combines this with more recent developments in the field of domain independent planning such as distance based heuristics and reachability analysis. We present an adaptation of the additive heuristic for plan space planning, and modify it to account for possible reuse of existing actions in a plan. We also propose a large set of novel flaw selection strategies, and show how these can help us solve more problems than previously possible by POCL planners. VHPOP also supports planning with durative actions by incorporating standard techniques for temporal constraint reasoning. We demonstrate that the same heuristic techniques used to boost the performance of classical POCL planning can be effective in domains with durative actions as well. The result is a versatile heuristic POCL planner competitive with established CSP-based and heuristic state space planners.