Planning under uncertainty is a central problem in the study of automated sequential decision making, and has been addressed by researchers in many different fields, including AI planning, decision analysis, operations research, control theory and economics. While the assumptions and perspectives adopted in these areas often differ in substantial ways, many planning problems of interest to researchers in these fields can be modeled as Markov decision processes (MDPs) and analyzed using the techniques of decision theory. This paper presents an overview and synthesis of MDP-related methods, showing how they provide a unifying framework for modeling many classes of planning problems studied in AI. It also describes structural properties of MDPs that, when exhibited by particular classes of problems, can be exploited in the construction of optimal or approximately optimal policies or plans. Planning problems commonly possess structure in the reward and value functions used to describe performance criteria, in the functions used to describe state transitions and observations, and in the relationships among features used to describe states, actions, rewards, and observations. Specialized representations, and algorithms employing these representations, can achieve computational leverage by exploiting these various forms of structure. Certain AI techniques -- in particular those based on the use of structured, intensional representations -- can be viewed in this way. This paper surveys several types of representations for both classical and decision-theoretic planning problems, and planning algorithms that exploit these representations in a number of different ways to ease the computational burden of constructing policies or plans. It focuses primarily on abstraction, aggregation and decomposition techniques based on AI-style representations.

As planning is applied to larger and richer domains the effort involved in constructing domain descriptions increases and becomes a significant burden on the human application designer. If general planners are to be applied successfully to large and complex domains it is necessary to provide the domain designer with some assistance in building correctly encoded domains. One way of doing this is to provide domain-independent techniques for extracting, from a domain description, knowledge that is implicit in that description and that can assist domain designers in debugging domain descriptions. This knowledge can also be exploited to improve the performance of planners: several researchers have explored the potential of state invariants in speeding up the performance of domain-independent planners. In this paper we describe a process by which state invariants can be extracted from the automatically inferred type structure of a domain. These techniques are being developed for exploitation by STAN, a Graphplan based planner that employs state analysis techniques to enhance its performance.

Stacked generalization is a general method of using a high-level model to combine lower-level models to achieve greater predictive accuracy. In this paper we address two crucial issues which have been considered to be a `black art' in classification tasks ever since the introduction of stacked generalization in 1992 by Wolpert: the type of generalizer that is suitable to derive the higher-level model, and the kind of attributes that should be used as its input. We find that best results are obtained when the higher-level model combines the confidence (and not just the predictions) of the lower-level ones. We demonstrate the effectiveness of stacked generalization for combining three different types of learning algorithms for classification tasks. We also compare the performance of stacked generalization with majority vote and published results of arcing and bagging.

A class of interval-based temporal languages for uniformly representing and reasoning about actions and plans is presented. Actions are represented by describing what is true while the action itself is occurring, and plans are constructed by temporally relating actions and world states. The temporal languages are members of the family of Description Logics, which are characterized by high expressivity combined with good computational properties. The subsumption problem for a class of temporal Description Logics is investigated and sound and complete decision procedures are given. The basic language TL-F is considered first: it is the composition of a temporal logic TL -- able to express interval temporal networks -- together with the non-temporal logic F -- a Feature Description Logic. It is proven that subsumption in this language is an NP-complete problem. Then it is shown how to reason with the more expressive languages TLU-FU and TL-ALCF. The former adds disjunction both at the temporal and non-temporal sides of the language, the latter extends the non-temporal side with set-valued features (i.e., roles) and a propositionally complete language.

We describe a general approach to optimization which we term `Squeaky Wheel' Optimization (SWO). In SWO, a greedy algorithm is used to construct a solution which is then analyzed to find the trouble spots, i.e., those elements, that, if improved, are likely to improve the objective function score. The results of the analysis are used to generate new priorities that determine the order in which the greedy algorithm constructs the next solution. This Construct/Analyze/Prioritize cycle continues until some limit is reached, or an acceptable solution is found. SWO can be viewed as operating on two search spaces: solutions and prioritizations. Successive solutions are only indirectly related, via the re-prioritization that results from analyzing the prior solution. Similarly, successive prioritizations are generated by constructing and analyzing solutions. This `coupled search' has some interesting properties, which we discuss. We report encouraging experimental results on two domains, scheduling problems that arise in fiber-optic cable manufacturing, and graph coloring problems. The fact that these domains are very different supports our claim that SWO is a general technique for optimization.

Many of the artificial intelligence techniques developed to date rely on heuristic search through large spaces. Unfortunately, the size of these spaces and the corresponding computational effort reduce the applicability of otherwise novel and effective algorithms. A number of parallel and distributed approaches to search have considerably improved the performance of the search process. Our goal is to develop an architecture that automatically selects parallel search strategies for optimal performance on a variety of search problems. In this paper we describe one such architecture realized in the Eureka system, which combines the benefits of many different approaches to parallel heuristic search. Through empirical and theoretical analyses we observe that features of the problem space directly affect the choice of optimal parallel search strategy. We then employ machine learning techniques to select the optimal parallel search strategy for a given problem space. When a new search task is input to the system, Eureka uses features describing the search space and the chosen architecture to automatically select the appropriate search strategy. Eureka has been tested on a MIMD parallel processor, a distributed network of workstations, and a single workstation using multithreading. Results generated from fifteen puzzle problems, robot arm motion problems, artificial search spaces, and planning problems indicate that Eureka outperforms any of the tested strategies used exclusively for all problem instances and is able to greatly reduce the search time for these applications.

This paper offers an approach to extensible knowledge representation and reasoning for a family of formalisms known as Description Logics. The approach is based on the notion of adding new concept constructors, and includes a heuristic methodology for specifying the desired extensions, as well as a modularized software architecture that supports implementing extensions. The architecture detailed here falls in the normalize-compared paradigm, and supports both intentional reasoning (subsumption) involving concepts, and extensional reasoning involving individuals after incremental updates to the knowledge base. The resulting approach can be used to extend the reasoner with specialized notions that are motivated by specific problems or application areas, such as reasoning about dates, plans, etc. In addition, it provides an opportunity to implement constructors that are not currently yet sufficiently well understood theoretically, but are needed in practice. Also, for constructors that are provably hard to reason with (e.g., ones whose presence would lead to undecidability), it allows the implementation of incomplete reasoners where the incompleteness is tailored to be acceptable for the application at hand.

It is common to view programs as a combination of logic and control: the logic part defines what the program must do, the control part -- how to do it. The Logic Programming paradigm was developed with the intention of separating the logic from the control. Recently, extensive research has been conducted on automatic generation of control for logic programs. Only a few of these works considered the issue of automatic generation of control for improving the efficiency of logic programs. In this paper we present a novel algorithm for automatic finding of lowest-cost subgoal orderings. The algorithm works using the divide-and-conquer strategy. The given set of subgoals is partitioned into smaller sets, based on co-occurrence of free variables. The subsets are ordered recursively and merged, yielding a provably optimal order. We experimentally demonstrate the utility of the algorithm by testing it in several domains, and discuss the possibilities of its cooperation with other existing methods.

We study the problem of probabilistic deduction with conditional constraints over basic events. We show that globally complete probabilistic deduction with conditional constraints over basic events is NP-hard. We then concentrate on the special case of probabilistic deduction in conditional constraint trees. We elaborate very efficient techniques for globally complete probabilistic deduction. In detail, for conditional constraint trees with point probabilities, we present a local approach to globally complete probabilistic deduction, which runs in linear time in the size of the conditional constraint trees. For conditional constraint trees with interval probabilities, we show that globally complete probabilistic deduction can be done in a global approach by solving nonlinear programs. We show how these nonlinear programs can be transformed into equivalent linear programs, which are solvable in polynomial time in the size of the conditional constraint trees.

Order of magnitude reasoning - reasoning by rough comparisons of the sizes of quantities - is often called 'back of the envelope calculation', with the implication that the calculations are quick though approximate. This paper exhibits an interesting class of constraint sets in which order of magnitude reasoning is demonstrably fast. Specifically, we present a polynomial-time algorithm that can solve a set of constraints of the form 'Points a and b are much closer together than points c and d.' We prove that this algorithm can be applied if `much closer together' is interpreted either as referring to an infinite difference in scale or as referring to a finite difference in scale, as long as the difference in scale is greater than the number of variables in the constraint set. We also prove that the first-order theory over such constraints is decidable.