By Eugene Fink

The function of our study is to augment the potency of AI challenge solvers via automating illustration adjustments. we've constructed a approach that improves the outline of enter difficulties and selects a suitable seek set of rules for every given challenge. Motivation. Researchers have accrued a lot proof at the impor­ tance of applicable representations for the potency of AI structures. an identical challenge might be effortless or tricky, reckoning on the way in which we describe it and at the seek set of rules we use. prior paintings at the computerized im­ provement of challenge descriptions has typically been constrained to the layout of person studying algorithms. The person has ordinarily been chargeable for the alternative of algorithms applicable for a given challenge. We current a procedure that integrates a number of description-changing and problem-solving algorithms. the aim of the said paintings is to formalize the concept that of illustration and to verify the next speculation: a good representation-changing method should be outfitted from 3 elements: • a library of problem-solving algorithms; • a library of algorithms that enhance challenge descriptions; • a keep watch over module that selects algorithms for every given problem.

Show description

Read Online or Download Changes of Problem Representation: Theory and Experiments PDF

Similar algorithms books

Algorithms For Interviews

Algorithms For Interviews (AFI) goals to assist engineers interviewing for software program improvement positions in addition to their interviewers. AFI includes 174 solved set of rules layout difficulties. It covers middle fabric, resembling looking out and sorting; basic layout ideas, similar to graph modeling and dynamic programming; complex subject matters, equivalent to strings, parallelism and intractability.

Scalable Optimization via Probabilistic Modeling: From Algorithms to Applications (Studies in Computational Intelligence, Volume 33)

This booklet focuses like a laser beam on one of many most well liked themes in evolutionary computation over the past decade or so: estimation of distribution algorithms (EDAs). EDAs are a massive present strategy that's resulting in breakthroughs in genetic and evolutionary computation and in optimization extra ordinarily.

Abstract Compositional Analysis of Iterated Relations: A Structural Approach to Complex State Transition Systems

This self-contained monograph is an built-in learn of usual structures outlined by means of iterated family members utilizing the 2 paradigms of abstraction and composition. This comprises the complexity of a few state-transition structures and improves figuring out of complicated or chaotic phenomena rising in a few dynamical structures.

Estimation of Distribution Algorithms: A New Tool for Evolutionary Computation

Estimation of Distribution Algorithms: a brand new device for Evolutionary Computation is dedicated to a brand new paradigm for evolutionary computation, named estimation of distribution algorithms (EDAs). This new classification of algorithms generalizes genetic algorithms through exchanging the crossover and mutation operators with studying and sampling from the likelihood distribution of the simplest participants of the inhabitants at every one generation of the set of rules.

Extra info for Changes of Problem Representation: Theory and Experiments

Example text

1998] have begun development of an architecture for integration of AI planning techniques. Their architecture, called PLAN++, includes tools for implementing, modifying, and reusing the main elements of planning systems. The purpose is to modularize typical planning algorithms, construct a large library of search modules, and use them as building blocks for new algorithms. Since the effectiveness of planning techniques varies across domains, the authors of PLAN++ intend to design software tools that enable the user to select appropriate modules and configure them for specific domains.

If this information is available, the changer algorithm utilizes it to generate a better description; otherwise, the algorithm uses some default assumptions. The optional input may include restrictions on the allowed problem instances, useful knowledge about domain properties, and advice from the user. For example, we may specify constraints on the allowed problems as an input to Margie. 2). As another example, the user may pre-select some primary effects of operators. 1). 14; however, this specification does not account for advanced features of the PRODIGY domain language.

3, we will describe a general utility function that unifies the three evaluation factors. When evaluating the utility of a new representation, we have to account for the overall time for improving a domain description, selecting a problem solver, and using the resulting representation to solve given problems. The system is effective only if this overall time is smaller than the time for solving the problems with the initial description and some fixed solver algorithm. 2 Specifications of description changers The current version of SHAPER includes seven description changers.

Download PDF sample

Rated 4.75 of 5 – based on 23 votes