This paper presents Newton, a branch & prune algorithm to find all isolated solutions of a system of polynomial constraints. Newton can be characterized as a global search method which uses intervals for numerical correctness and for pruning the search space early. The pruning in Newton consists in enforcing at each node of the search tree a unique local consistency condition, called box-consistency, which approximates the notion of arc-consistency well-known in artificial intelligence. Box-consistency is parametrized by an interval extension of the constraint and can be instantiated to produce the Hansen-Segupta's narrowing operator (used in interval methods) as well as new operators which are more effective when the computation is far from a solution. Newton has been evaluated on a variety of benchmarks from kinematics, chemistry, combustion, economics, and mechanics. On these benchmarks, it outperforms the interval methods we are aware of and compares well with state-of-the-art continuation methods. Limitations of Newton (e.g., a sensitivity to the size of the initial intervals on some problems) are also discussed. Of particular interest is the mathematical and programming simplicity of the method.

It has been argued that the use of constraint-based techniques and tools enables the implementation of precise, flexible, efficient and extensible scheduling systems: precise and flexible as the system can take into account any constraint expressible in the constraint language; efficient inasmuch as highly optimized constraint propagation procedures are now available; extensible as the consideration of a new type of constraint may require (especially in an object-oriented framework) only an extension to the constraint system or, in the worst case, the implementation of additional decision-making modules (without needs for modification of the existing code). The following paper presents ILOG SCHEDULE, a C++ library enabling the representation of a wide collection of scheduling constraints in terms of "resources" and "activities." ILOG SCHEDULE is based on SOLVER, the generic software tool for object-oriented constraint programming developed and marketed by ILOG. SOLVER variables and constraints can be accessed from SCHEDULE activities and resources. As a result, the user of SCHEDULE can make use of SOLVER to represent specific constraints, and implement and combine the specific problem-solving strategies that are the most appropriate for the scheduling application under consideration. It is hoped --- and expected --- that object-oriented constraint programming tools like SCHEDULE will enable the industry to make decisive steps toward the implementation of "state of the art," highly flexible, constraint-based scheduling applications.

Shortcomings of qualitative simulation and of quantitative simulation motivate combining them to do simulations exhibiting strengths of both. The resulting class of techniques is called semi-quantitative simulation. One approach to semi-quantitative simulation is to use numeric intervals to represent incomplete quantitative information. In this research we demonstrate semiquantitative simulation using intervals in an implemented semi-quantitative simulator called Q3. Q3 progressively refines a qualitative simulation, providing increasingly specific quantitative predictions which can converge to a numerical simulation in the limit while retaining important correctness guarantees from qualitative and interval simulation techniques. Q3's simulations are based on a technique we call step size refinement. While a pure qualitative simulation has a very coarse step size, representing the state of a system trajectory at relatively few qualitatively distinct states, Q3 interpolates newly expl...

Nondeterministic Lisp is a simple extension of Lisp which provides automatic backtracking. Nondeterminism allows concise description of many search tasks which form the basis of much AI research. This paper discusses Screamer, an efficient implementation of nondeterministic Lisp as a fully portable extension of Common Lisp. In this paper we present the basic nondeterministic Lisp constructs, motivate the utility of the language via numerous short examples, and discuss the compilation techniques. Supported in part by an AT&T Bell Laboratories Ph.D. scholarship to the author, by a Presidential Young Investigator Award to Professor Robert C. Berwick under National Science Foundation Grant DCR-- 85552543, by a grant from the Siemens Corporation, and by the Kapor Family Foundation. Also supported in part by ARO grant DAAL 03--89--C--0031, by DARPA grant N00014--90--J--1863, by NSF grant IRI 90-- 16592, and by Ben Franklin grant 91S.3078C--1 y Supported in part by the Advanced Resea...

In this paper we first question whether control and initiative are as synchronous as current usages appear to make them. We provide some evidence from a corpus of collaborative problem solving dialogues that control should apply to the dialogue level and initiative to the problem solving level. We then ask what role these two phenomena have in mixed initiative dialogues and present an example of how the recognition of lack of initiative at the problem solving level is cued by control at the dialogue level. Introduction In this brief position paper, we would like to pose two questions we hope are of interest to the symposium participants: 1. Are control and initiative the same phenomena? We provide some evidence that they are not. 2. What is the importance of such a distinction for mixed initiative interactive dialogues? We present some preliminary thoughts on this issue. Control vs. Initiative Control and initiative are two terms that appear to be used almost as synonyms in the li...

Computer programming environments for learning should make it easy to create worlds of responsive and autonomous objects, such as video games or simulations of animal behavior. But building such worlds remains difficult, partly because the models and metaphors underlying traditional programming languages are not particularly suited to the task. This dissertation investigates new metaphors, environments, and languages that make possible new ways to create programs -- and, more broadly, new ways to think about programs. In particular, it introduces the idea of programming with "agents" as a means to help people create worlds involving responsive, interacting objects. In this context, an agent is a simple mechanism intended to be understood through anthropomorphic metaphors and endowed with certain lifelike properties such as autonomy, purposefulness, and emotional state. Complex behavior is achieved by combining simple agents into more complex structures. While the agent metaphor enables...

We provide evidence that speech act recognition, is 1) difficult for humans to do and 2) likely to misidentify proposals involving reconstructed intentions. We examine the reliability of coding for speech acts in collaborative dialogues and we present an approach for recognizing reconstructed proposals using domain context and other more easily recognized features. 1 Introduction Speech act recognition plays a prominent role in dialogue understanding, in traditional approaches that infer a plan using plan construction operators [PA80], [LA90], [LC91, LC92], and in more recent techniques relying on statistical correlations or finite state machines [RM95, QDL + 97]. Both approaches recognize surface speech acts, using surface form and information provided by the discourse context and the discourse operators, or by a finite state approximation of the planning information. These approaches assume that it is (relatively) simple to recognize speech acts, and that speech acts are a requi...

Abstract not found

ion Projection rule 1 is sound, i.e., it does not leave out any possible "post-execution" probability distribution. Furthermore project 1 (a; M pre ) can be computed much faster than exec(a; M pre ). Unfortunately, in plan projection, even project 1 is not fast enough, since its computational time grows exponentially as a function of plan length. We would like to seek project 2 and action a (a as a function of a) such that M post satisfying M post = project 2 (a ; M pre ) and M post ' exec(a; M pre ) could be computed faster than M post . As we observed, projecting action a on M pre amounts to projecting every branch (c l ; I l ; E l ) of a on every branch of M pre . We conclude therefore that if a has fewer branches than a then projecting a should be faster. This can be done by abstracting a set of branches of a into a single branch of a . We refer to this process as intra-action abstraction. Given two actions a 1 and a 2 we could seek project 2 and a...

FaCiLe is an open source constraint programming library over integer nite domain written in OCaml, a functional language of the ML family. It oers all usual constraint system facilities to create and handle nite domain variables, arithmetic constraints (possibly nonlinear) , built-in global constraints and search goals. FaCiLe allows as well to build easily user-dened constraints and goals from scratch or by combining simple primitives, making pervasive use of higher-order functionals to provide a simple and exible user interface. As FaCiLe is an OCaml library and not yet another language, the user benets from polymorphic type inference and strong typing discipline, high level of abstraction, generic modules and object system, as well as native code compilation eciency, garbage collection and replay debugger. All these features allow to prototype and experiment quickly: modelling, data processing and interface are implemented in the same powerful language with a high level of safety.