Optimization Techniques for Solving Complex Problems

CHAPTER 10

Genetic Algorithms, Parallelism, and Reconfigurable Hardware

J. M. SÁNCHEZ

Universidad de Extremadura, Spain

M. RUBIO

Centro Extremeño de Tecnologías Avanzadas, Spain

M. A. VEGA and J. A. GÓMEZ

Universidad de Extremadura, Spain

10.1 INTRODUCTION

The grid computing model offers a powerful and high-availability platform that allows efficient running of applications that are intensive in terms of calculations and data. It is a distributed model based on the coordinated use of several spread management services and computational resources. Its main characteristic is resource sharing; collaboration is the philosophy. New nodes can easily aggregate to an organization (named V.O., virtual organization, in grid terms), independent of its geographical situation. All resources shared within the V.O. will conform to a federated resource available for any member of it. That is, the EELA project [1] gathers power from sites placed in both Latin American and Europe, and any member of it can benefit from its aggregated resources. This gives the grid model the power to establish a resource with virtually unlimited computing power. The grid computing model is the best suitable for applications whose throughput can be increased by multiplying the executions per unit time (e.g., parametric jobs, which perform huge amounts of similar calculations just varying the input parameters). Grid ...