Reactive search is developed around the metaphor of a chemical reaction empowering other meta-heuristic methods with more intelligent search functions. This happens in two ways: (1) by automatically speeding up the reactive process that extends the parameters of the solution possibilities (i.e., widening the solutions probabilities) and by intelligently refining the solution to a more "finely tuned" problem solution. Hence, it produces solutions that are drawn from a wider sphere of possibilities and less approximated than tabu search and other meta-heuristic strategies. The book presents the main principles of reactive search and intelligent optimization and clearly shows how they can be used in problem solving. It is organized to first introduce the basic issues andalgorithms, then identify the parameters critical for successfully using the methods discussed, and, finally, to present opportunities and schemes for the automated tuning of those parameters. Reactive Search and Intelligent Optimization, and meta-heuristics in general, support managers in decision-making withrobust tools that provide high-quality solutions to important applications inbusiness, engineering, economics and science in reasonable time horizons Especially when we think about systems development, meta-heuristics gain more and more acceptance in real-world applications such as ERP systems INDICE: Preface.- Introduction.- Reacting on the neighborhood.- Reacting on the annealing schedule.- Reactive prohibitions.- Model-based search.- Reacting on the objective function.- Reinforcement learning.- Algorithm portfolios and restart strategies.- Racing.- Metrics, landscapes, and features.- Relationships between reactive search and reinforcement learning.- Index.
- ISBN: 978-0-387-09623-0
- Editorial: Springer
- Encuadernacion: Cartoné
- Páginas: 190
- Fecha Publicación: 01/10/2008
- Nº Volúmenes: 1
- Idioma: Inglés