Evolutionary algorithms to face computer systems management problems

• Autores: Gallard, Raúl Hector; Esquivel, Susana Cecilia
• Año de publicación: 2000
• Idioma: inglés
• Tipo de recurso: documento de conferencia
• Materia: Ciencias Informáticas; evolutionary algorithms; resource management; computer systems; task allocation
• Institución: Universidad Nacional de La Plata

Evolutionary Algorithms (EAs) became a powerful tool for environments where an optimization is needed or where an adaptive behaviour of a system is necessary. Computer Systems have been evolving rapidly during the last five decades. Since the 70’s, due to improvements in communication and computer hardware, an evolution of computer systems based on networked workstations has been taking place. This advance triggered present distributed systems. At each stage of this evolution, resource management was designer’s main concern. Decisions on how to allocate a limited set of expensive resources and how to schedule arriving tasks with diverse requirements were based on various heuristics, attempting optimization of, frequently conflicting objectives, from the user and system perspectives. This paper discusses feasible areas where EAs can be efficiently applied to solve resource management problems in Computer systems, show two feasible implementations and introduces last improvements to enhance EAs performance
I Workshop de Agentes y Sistemas Inteligentes (WASI)
Red de Universidades con Carreras en Informática (RedUNCI)