We propose a definition of `constrainedness' that unifies two of the most common but informal uses of the term. These are that branching heuristics in search algorithms often try to make the most "constrained" choice, and that hard search problems tend to be "critically constrained". Our definition of constrainedness generalizes a number of parameters used to study phase transition behaviour in a wide variety of problem domains. As well as predicting the location of phase transitions in solubility, constrainedness provides insight into why problems at phase transitions tend to be hard to solve. Such problems are on a constrainedness "knife-edge", and we must search deep into the problem before they look more or less soluble. Heuristics that try to get off this knife-edge as quickly as possible by, for example, minimizing the constrainedness are often very effective. We show that heuristics from a wide variety of problem domains can be seen as minimizing the constrainedness (or proxies ...

The long-term goal of our field is the creation and understanding of intelligence. Productive research in AI, both practical and theoretical, benefits from a notion of intelligence that is precise enough to allow the cumulative development of robust systems and general results. This paper outlines a gradual evolution in our formal conception of intelligence that brings it closer to our informal conception and simultaneously reduces the gap between theory and practice. 1 Artificial Intelligence AI is a field in which the ultimate goal has often been somewhat ill-defined and subject to dispute. Some researchers aim to emulate human cognition, others aim at the creation of

Abstract not found

. In this paper we describe SINERGY, which is a highly parallelizable, linear planning system that is based on the genetic programming paradigm. Rather than reasoning about the world it is planning for, SINERGY uses artificial selection, recombination and fitness measure to generate linear plans that solve conjunctive goals. We ran SINERGY on several domains (e.g., the briefcase problem and a few variants of the robot navigation problem), and the experimental results show that our planner is capable of handling problem instances that are one to two orders of magnitude larger than the ones solved by UCPOP. In order to facilitate the search reduction and to enhance the expressive power of SINERGY, we also propose two major extensions to our planning system: a formalism for using hierarchical planning operators, and a framework for planning in dynamic environments. 1 Motivation Artificial intelligence planning is a notoriously hard problem. There are several papers [Chapman 1987, Joslin...

Most conventional law equation discovery systems suchasBACON require experimental environments to acquire their necessary data. The mathematical techniques such as linear system identification and neural network fitting presume the classes of equations to model given observed data sets. The study reported in this paper proposes a novel method to discover an admissible model equation from a given set of observed data, while the equation is ensured to reflect first principles governing the objective system. The power of the proposed method comes from the use of the scale-types of the observed quantities, a mathematical property of identity and quasi-bi-variate fitting to the given data set. Its principles and algorithm are described with moderately complex examples, and its practicality is demonstrated through a real application to psychological and sociological law equation discovery. 1 Introduction The most well known pioneering system to discover scientific law eq...

In this paper we describe SYNERGY, which is a general-purpose AI planning system that is based on the genetic programming paradigm. Rather than reasoning about the planning world, SYNERGY uses selection, mutation, recombination and fitness measure to generate linear plans that solve conjunctive goals. We ran SYNERGY on several domains, and the experimental results show that our planner solves problem instances that are up to two orders of magnitude larger than the ones solved by UCPOP. KEYWORDS: AI planning, genetic programming, conjunctive goals INTRODUCTION Artificial intelligence (AI) planning is known to be an extremely hard problem (see [2]), and it is generally accepted that most non-trivial planning problems are at least NPcomplete. In order to cope with the combinatorial explosion of the search problem, AI researchers proposed a wide variety of solutions, from search control rules [3] to hierarchical planning [8] to skeletal planning [5]. More recently, we witnessed the occur...

This paper describes a program called Pret that automates system identication, the process of nding a dynamical model of a black-box system. Pret performs both structural identication and parameter estimation by integrating several reasoning modes: qualitative reasoning, qualitative simulation, numerical simulation, geometric reasoning, constraint reasoning, resolution, reasoning with abstraction levels, declarative meta-level control, and a simple form of truth maintenance. Unlike other modeling programs that map structural or functional descriptions to model fragments, Pret combines hypotheses about the mathematics involved into candidate models that are intelligently tested against observations about the target system. We give two examples of system identication tasks that this automated modeling tool has successfully performed. The rst, a simple linear system, was chosen because it facilitates a brief and clear presentation of Pret's features and reasoning t...

this paper, we attempt to address its efficacy. In particular, we would like to address the issues of (a) whether this reactive approach is viable (b) what advantages, if any, it provides and (c) where it causes problems. The discussion is made relative to the XII implementation as embedded in the Softbot called Rodney. By doing so, we hope to elucidate the trade-offs involved in choosing between the two approaches.

. When a metaprogram automatically creates rules, some created rules are useless because they can never apply. Some metarules, that we call impossibility metarules, are used to remove useless rules. Some of these metarules are general and apply to any generated program. Some are domain specific metarules. In this paper, we show how dynamic metaprogramming can be used to create domain specific impossibility metarules. Applying metaprogramming to impossibility metaprogramming avoids writing specific metaprogram for each domain metaprogramming is applied to. Our metametaprograms have been used to write metaprograms that write search rules for different games and planning domains. They write programs that write selective and efficient search programs. 1 Introduction Knowledge about the moves to try enables to select a small number of moves from a possibly large set of possible moves. It is very important in complex games and planning domains where search trees have a large bran...

We present a logic programming system that accomplishes three important goals: equivalence of declarative and operational semantics, declarative specification of control information, and smoothness of interaction with nonlogic -based programs. The language of the system is that of Generalized Horn Clause Intuitionistic Logic with negation as inconsistency. Meta level predicates are used to specify control information declaratively, compensating for the absence of procedural constructs that usually facilitate formulation of e#- cient programs. Knowledge that has been derived in the course of the current inference process can at any time be passed to non-logic-based program modules. Traditional SLD inference engines maintain only the linear path to the current state in the SLD search tree: formulae that have been proved on this path are implicitly represented in a stack of recursive calls to the inference engine, and formulae that have been proved on previous, unsuccessful paths are lost...