Jana Koehler

Encoding Planning Problems in Nonmonotonic Logic Programs

We present a framework for encoding planning problems in logic programs with negation as failure, having computational efficiency as our major consideration. In order to accomplish our goal, we bring together ideas from logic programming and the planning systems GRAPHPLAN and SATPLAN. We discuss different representations of planning problems in logic programs, point out issues related to their performance, and show ways to exploit the structure of the domains in these representations. For our experimentation we use an existing implementation of the stable models semantics called SMODELS. It turns out that for careful and compact encodings, the performance of the method across a number of different domains, is comparable to that of planners like GRAPHPLAN and SATPLAN.

On reasonable and forced goal orderings and their use in an agenda-driven planning algorithm

The paper addresses the problem of computing goal orderings, which is one of the longstanding issues in AI planning. It makes two new contributions. First, it formally defines and discusses two different goal orderings, which are called the reasonable and the forced ordering. Both orderings are defined for simple STRIPS operators as well as for more complex ADL operators supporting negation and conditional effects. The complexity of these orderings is investigated and their practical relevance is discussed. Secondly, two different methods to compute reasonable goal orderings are developed. One of them is based on planning graphs, while the other investigates the set of actions directly. Finally, it is shown how the ordering relations, which have been derived for a given set of goals G, can be used to compute a so-called goal agenda that divides G into an ordered set of subgoals. Any planner can then, in principle, use the goal agenda to plan for increasing sets of subgoals. This can lead to an exponential complexity reduction, as the solution to a complex planning problem is found by solving easier subproblems. Since only a polynomial overhead is caused by the goal agenda computation, a potential exists to dramatically speed up planning algorithms as we demonstrate in the empirical evaluation, where we use this method in the IPP planner.

Ignoring Irrelevant Facts and Operators in Plan Generation

It is traditional wisdom that one should start from the goals when generating a plan in order to focus the plan generation process on potentially relevant actions. The GRAPHPLAN system, however, which is the most efficient planning system nowadays, builds a “planning graph” in a forward-chaining manner. Although this strategy seems to work well, it may possibly lead to problems if the planning task description contains irrelevant information. Although some irrelevant information can be filtered out by GRAPHPLAN, most cases of irrelevance are not noticed.

An AI-based approach to destination control in elevators

Not widely known by the AI community, elevator control has become a major field of application for AI technologies. Techniques such as neural networks, genetic algorithms, fuzzy rules and, recently, multiagent systems and AI planning have been adopted by leading elevator companies not only to improve the transportation capacity of conventional elevator systems but also to revolutionize the way in which elevators interact with and serve passengers. In this article, we begin with an overview of AI techniques adopted by this industry and explain the motivations behind the continuous interest in AI. We review and summarize publications that are not easily accessible from the common AI sources. In the second part, we present in more detail a recent development project to apply AI planning and multiagent systems to elevator control problems.

Modal logics, description logics and arithmetic reasoning

Forthcoming in the Journal of Ariticial Intelligence We introduce mathematical programming and atomic decomposition as the basic modal (T-Box) inference techniques for a large class of modal and description logics. The class of description logics suitable for the proposed methods is strong on the arithmetical side. In particular there may be complex arithmetical conditions on sets of accessible worlds (role llers). The atomic decomposition technique can deal with set constructors for modal parameters (role terms) and parameter (role) hierarchies specied in full propositional logic. Besides the standard modal operators, a number of other constructors can be added in a relatively straightforward way. Examples are graded modalities (qualied number restrictions) and also generalized quantiers like ‘most’, ‘n%’, ‘more’ and ‘many’.

The AIPS-98 Planning Competition

In 1998, the international planning community was invited to take part in the first planning competition, hosted by the Artificial Intelligence Planning Systems Conference, to provide a new impetus for empirical evaluation and direct comparison of automatic domain-independent planning systems. This article describes the systems that competed in the event, examines the results, and considers some of the implications for the future of the field.

Planning from second principles

Abstract Planning from second principles by reusing and modifying plans is one way of improving the efficiency of planning systems. In this paper, we study it in the general framework of deductive planning and develop a logical formalization of planning from second principles, which relies on a systematic decomposition of the planning process. Deductive inference processes with clearly defined semantics formalize each of the subtasks a second principles planner has to address. Plan modification, which comprises matching and adaptation tasks, is based on a deductive approach yielding provably correct modified plans. Description logics are introduced as query languages to plan libraries, which leads to a novel and efficient solution to the indexing problem in case-based reasoning. Apart from sequential plans, this approach enables a planner to reuse and modify complex plans containing control structures like conditionals and loops.

An Application of Terminological Logics to Case-based Reasoning.

Abstract A key problem in case-based reasoning is the representation, organization and maintenance of case libraries. While current approaches rely on heuristic and psychologically inspired formalisms, terminological logics have emerged as a powerful representation formalism with clearly defined formal semantics. This paper demonstrates how the indexing of case libraries can be grounded on terminological logics by using them as a kind of query language to the case library. Indices of cases are represented as concepts in a terminological logic. They are automatically constructed from the symbolic representation of cases with the help of a well-defined abstraction process. The retrieval of cases from the library is grounded on concept classification. The theoretical approach provides the formal foundation for the fully implemented case-based planning system MRL. The use of terminological logics allows formal proof of properties like the correctness, completeness and efficiency of the retrieval algorithm, which has rarely been done for existing case-based reasoning systems.

Deductive planning and plan reuse in a command language environment

In this paper we introduce a deductive planning system currently being developed as the kernel of an intelligent help system. It consists of a deductive planner and a plan reuse component and with that provides planning from first as well as planning from second principles. Both components rely upon an interval-based temporal logic. The deductive formalisms realizing plan formation from formal specifications and the reuse of already existing plans respectively are presented and demonstrated by examples taken from an operating systems domain.

Constraint Deduction in an Interval-based Temporal Logic

We describe reasoning methods for the interval-based modal temporal logic LLP which employs the modal operators sometimes, always, next, and chop. We propose a constraint deduction approach and compare it with a sequent calculus, developed as the basic machinery for the deductive planning system PHI which uses LLP as underlying formalism.