Classical planning is the problem of finding a sequence of actions to achieve a goal given an exact characterization of a domain. An algorithm to solve this problem is presented, which searches a space of plan prefixes, trying to extend one of them to a complete sequence of actions. It is guided by a heuristic estimator based on regression-match graphs, which attempt to characterize the entire subgoal structure of the remaining part of the problem. These graphs simplify the structure by neglecting goal interactions and by assuming that variables in goal conjunctions should be bound in such a way as to make as many conjuncts as possible true without further work. In some domains, these approximations work very well, and experiments show that many classical-planning problems can solved with very little search. 1 Definition of the Problem The classical planning problem is to generate a sequence of actions that make a given proposition true, in a domain in which there is perfect informati...

. Humans have different problem solving strategies at their disposal and they can flexibly employ several strategies when solving a complex problem, whereas previous theorem proving and planning systems typically employ a single strategy or a hard coded combination of a few strategies. We introduce multi-strategy proof planning that allows for combining a number of strategies and for switching flexibly between strategies in a proof planning process. Thereby proof planning becomes more robust since it does not necessarily fail if one problem solving mechanism fails. Rather it can reason about preference of strategies and about failures. Moreover, our strategies provide a means for structuring the vast amount of knowledge such that the planner can cope with the otherwise overwhelming knowledge in mathematics. 1 Introduction The choice of an appropriate problem solving strategy is a crucial human skill and is typically guided by some meta-level reasoning. Trained mathematicia...

This paper postulates a rigorous method for the construction of classical planning domain models. We describe, with the help of a non-trivial example, a tool supported method for encoding such models. The method results in an `object-centred' specification of the domain that lifts the representation from the level of the literal to the level of the object. Thus, for example, operators are defined in terms of how they change the state of objects, and planning states are defined as amalgams of the objects' states. The method features two classes of tools: for initial capture and validation of the domain model; and for operationalising the domain model (a process we call compilation) for later planning. Here we focus on compilation tools used to generate macros and goal orders to be utilised at plan generation time. We describe them in depth, and evaluate empirically their combined benefits in plan-generation speed-up. The method's main benefit is in helping the modeller to pro...

We are motivated by the problem of building agents that interact in complex real-world domains, such as UNIX and the Internet. Such agents must be able to exploit complete information when possible, yet cope with incomplete information when necessary. They need to distinguish actions that return information from those that change the world, and know when each type of action is appropriate. They must also be able to plan to obtain information needed for further planning. They should be able to represent and exploit the richness of their domains, including universally quantified causal (e.g., UNIX chmod *) and observational (e.g., ls) effects, which are ubiquitous in real-world domains such as the Internet. The xii planner solves the problems listed above by extending classical planner representations and algorithms to deal with incomplete information. xii represents and reasons about local closed world information, information preconditions and postconditions and universally quantified ...

Given the intractability of domain-independent planning, the ability to control the search of a planner is vitally important. One way of doing this involves learning from search failures. This paper describes SNLP+EBL, the first implementation of explanation based search control rule learning framework for a partial order (plan-space) planner. We will start by describing the basic learning framework of SNLP+EBL. We will then concentrate on SNLP+EBL's ability to learn from failures, and describe the results of empirical studies which demonstrate the effectiveness of the search-control rules SNLP+EBL learns using our method. We then

This thesis investigates problem-solving in domains featuring incomplete information and multiple agents with opposing goals. In particular, we describe Finesse --- a system that forms plans for the problem of declarer play in the game of Bridge. We begin by examining the problem of search. We formalise a best defence model of incomplete information games in which equilibrium point strategies can be identified, and identify two specific problems that can affect algorithms in such domains. In Bridge, we show that the best defence model corresponds to the typical model analysed in expert texts, and examine search algorithms which overcome the problems we have identified. Next, we look at how planning algorithms can be made to cope with the difficulties of such domains. This calls for the development of new techniques for representing uncertainty and actions with disjunctive effects, for coping with an opposition, and for reasoning about compound actions. We tackle these problems with a...

Hierarchical task network and action-based planning approaches have traditionally been studied separately. In many domains, human expertise in the form of hierarchical reduction schemas exists, but is incomplete. In such domains, hybrid approaches that use both HTN and action-based planning techniques are needed. In this paper, we extend our previous work on refinement planning to include hierarchical planning. Specifically, we provide a generalized plan-space refinement that is capable of handling non-primitive actions. The generalization provides a principled way of handling partially hierarchical domains, while preserving systematicity, and respecting the user-intent inherent in the reduction schemas. Our general account also puts into perspective the many surface differences between the HTN and action-based planners, and could support the transfer of progress between HTN and action-based planning approaches. 1 Introduction Traditionally, classical planning probl...

Planning and Knowledge Representation for Softbots by Keith Golden Chairperson of Supervisory Committee: Professor Dan Weld Computer Science and Engineering This thesis describes the design of a planner and knowledge representation languages for building software agents, known as softbots. While the focus of this thesis is on softbots, the ideas and algorithms presented here are general-purpose and could be applied to robotic domains as well. The major contributions are: ffl The lcw (Local Closed World) knowledge representation, used to capture an agent's incomplete information about the world, which can include localized closure information, such as knowledge of all files in a directory. We present lcw inference and update procedures that are sound, fast and effective. ffl The sadl action language, used to describe actions and goals available to the agent, including sensing actions and goals of acquiring new information. We define the semantics for sadl and we illustrate the exp...

Most existing planners intertwine the refinement of a partial plan with search by pushing the individual refinements of a plan into different search branches. Although this approach reduces the cost of handling partial plans, it also often leads to search space explosion. In this paper, we consider the possibility of handling the refinements of a partial plan together (without splitting them into search space). This is facilitated by disjunctive partial plan representations that can compactly represent large sets of partial plans. Disjunctive representations have hitherto been shunned since they may increase the plan handling costs. We argue that performance improvements can be obtained despite these costs by the use of (a) constraint propagation techniques to simplify the disjunctive plans and (b) CSP/SAT techniques to extract solutions from them. We will support this view by showing that some recent promising refinement planners, such as the GRAPHPLAN algorithm [2], can be seen as deriving their power from disjunctive plan representations. We will also present a new planning algorithm, UCPOPD, which uses disjunctive representations over UCPOP [19] to improve performance. Finally, we will discuss the issues and tradeoffs involved in planning with disjunctive representations.

Subgoal interactions have received considerable attention in AI Planning. Earlier analyses by Korf [11] and Joslin and Roach [6] were done in terms of the topology of the space of world states. More recent analyses by Barrett and Weld [1] and Veloso and Blythe [14] were done in terms of the nature of the planner. In this paper, we will argue that subgoal interactions are best understood in terms of the candidate sets of the plans for the individual subgoals. We will describe a generalized representation for partial plans that applies to a large class of refinement planners,and discuss the notion of mergeability and serial extensibility of these partial plans. The concepts of independence and serializability of subgoals are derived by generalizing mergeability and serial extensibility over classes of partial plans. Unlike previous work, our analysis also applies to multi-method refinement planners such as UCP [7]. We will show that all existing characterizations of serializability diffe...