Bernhard Nebel

The FF planning system: fast plan generation through heuristic search

We describe and evaluate the algorithmic techniques that are used in the FF planning system. Like the HSP system, FF relies on forward state space search, using a heuristic that estimates goal distances by ignoring delete lists. Unlike HSPs heuristic, our method does not assume facts to be independent. We introduce a novel search strategy that combines hill-climbing with systematic search, and we show how other powerful heuristic information can be extracted and used to prune the search space. FF was the most successful automatic planner at the recent AIPS-2000 planning competition. We review the results of the competition, give data for other benchmark domains, and investigate the reasons for the runtime performance of FF compared to HSP.

Reasoning about temporal relations: a maximal tractable subclass of Allen's interval algebra

We introduce a new subclass of Allens interval algebra we call “ORD-Horn subclass,” which is a strict superset of the “pointisable subclass.” We prove that reasoning in the ORD-Horn subclass is a polynomial-time problem and show that the path-consistency method is sufficient for deciding satisfiability. Further, using an extensive machine-generated case analysis, we show that the ORD-Horn subclass is a maximal tractable subclass of the full algebra (assuming P <inline-equation> <f> ≠</f> </inline-equation> NP). In fact, it is the unique greatest tractable subclass amongst the subclasses that contain all basic relations.

COMPLEXITY RESULTS FOR SAS+ PLANNING

We have previously reported a number of tractable planning problems defined in the SAS+ formalism. This article complements these results by providing a complete map over the complexity of SAS+ planning under all combinations of the previously considered restrictions. We analyze the complexity of both finding a minimal plan and finding any plan. In contrast to other complexity surveys of planning, we study not only the complexity of the decision problems but also the complexity of the generation problems. We prove that the SAS+‐PUS problem is the maximal tractable problem under the restrictions we have considered if we want to generate minimal plans. If we are satisfied with any plan, then we can generalize further to the SAS+‐US problem, which we prove to be the maximal tractable problem in this case.

Terminological reasoning is inherently intractable (research note)

Abstract Computational tractability has been a major concern in the area of terminological knowledge representation and reasoning. However, all analyses of the computational complexity of terminological reasoning are based on the hidden assumption that subsumption in terminologies reduces to subsumption of concept descriptions without a significant increase in computational complexity. In this paper it will be shown that this assumption, which seems to work in the “normal case,” is nevertheless wrong. Subsumption in terminologies turns out to be co-NP-complete for a minimal terminological representation language that is a subset of every useful terminological language.

An empirical analysis of optimization techniques for terminological representation systems : or: 'Making KRIS get a move on'

We consider different methods of optimizing the classification process of terminological representation systems, and evaluate their effect on three different types of test data. Though these techniques can probably be found in many existing systems, until now there has been no coherent description of these techniques and their impact on the performance of a system. One goal of this paper is to make such a description available for future implementors of terminological systems. Building the optimizations that came off best into the KRIS system greatly enhanced its efficiency.

Belief revision and default reasoning : syntax-based approaches

Belief revision leads to temporal nonmonotonicity, i.e., the set of beliefs does not grow monotonically with time. Default reasoning leads to logical nonmonotonicity, i.e., the set of consequences does not grow monotonically with the set of premises. The connection between these forms of nonmonotonicity will be studied in this paper focusing on syntax-based approaches. It is shown that a general form of syntax-based belief revision corresponds to a special kind of partial meet revision in the sense of variants of logics for default reasoning. Additionally, the computational complexity of the membership problem in revised belief sets and of the equivalent problem of derivability in default logics is analyzed, which turns out to be located at the lower end of the polynomial hierarchy.

Computational complexity of terminological reasoning in BACK

Abstract Terminological reasoning is a mode of reasoning all hybrid knowledge representation systems based on KL-ONE rely on. After a short introduction of what terminological reasoning amounts to, it is proven that a complete inference algorithm for the BACK system would be computationally intractable. Interestingly, this result also applies to the KANDOR system, which had been conjectured to realize complete terminological inferences with a tractable algorithm. More generally, together with an earlier paper of Brachman and Levesque it shows that terminological reasoning is intractable for any system using a nontrivial description language. Finally, consequences of this distressing result are briefly discussed.

TERMINOLOGICAL CYCLES: Semantics and Computational Properties

Abstract Terminological knowledge representation formalisms are intended to capture the analytic relationships between terms of a vocabulary intended to describe a domain. A term whose definition refers, either directly or indirectly, to the term itself presents a problem for most terminological representation systems because it is not obvious whether such a term is meaningful, nor how it could be handled by a knowledge representation system in a satisfying manner. After some examples of intuitively sound terminological cycles are given, different formal semantics are investigated and evaluated with respect to the examples. As it turns out, none of the different styles of semantics seems to be completely satisfying for all purposes. Finally, consequences in terms of computational complexity and decidability are discussed.

Spatial Cognition: Reasoning, Action, Interaction

The Transregional Collaborative Research Center SFB/TR 8 Spatial Cognition pursues interdisciplinary research on a broad range of topics related to the representation and processing mechanisms for intelligent spatial behavior in technical and in natural systems. This contribution gives an overview of the field of research worked on in the SFB/TR 8 Spatial Cognition and presents three representative examples that illustrate the activities in the three research areas Reasoning, Action, and Interaction.

Continual planning and acting in dynamic multiagent environments

In order to behave intelligently, artificial agents must be able to deliberatively plan their future actions. Unfortunately, realistic agent environments are usually highly dynamic and only partially observable, which makes planning computationally hard. For most practical purposes this rules out planning techniques that account for all possible contingencies in the planning process. However, many agent environments permit an alternative approach, namely continual planning, i.e. the interleaving of planning with acting and sensing. This paper presents a new principled approach to continual planning that describes why and when an agent should switch between planning and acting. The resulting continual planning algorithm enables agents to deliberately postpone parts of their planning process and instead actively gather missing information that is relevant for the later refinement of the plan. To this end, the algorithm explictly reasons about the knowledge (or lack thereof) of an agent and its sensory capabilities. These concepts are modelled in the planning language (MAPL). Since in many environments the major reason for dynamism is the behaviour of other agents, MAPL can also model multiagent environments, common knowledge among agents, and communicative actions between them. For Continual Planning, MAPL introduces the concept of of assertions, abstract actions that substitute yet unformed subplans. To evaluate our continual planning approach empirically we have developed MAPSIM, a simulation environment that automatically builds multiagent simulations from formal MAPL domains. Thus, agents can not only plan, but also execute their plans, perceive their environment, and interact with each other. Our experiments show that, using continual planning techniques, deliberate action planning can be used efficiently even in complex multiagent environments.