Designing Parallel Meta-Heuristic Methods

Abstract

Meta-heuristics represent powerful tools for addressing hard combinatorial optimization problems. However, real life instances usually cannot be treated efficiently in “reasonable” computing times. Moreover, a major issue in meta-heuristic design and calibration is to provide high performance solutions for a variety of problems. Parallel meta-heuristics aim to address both issues. The objective of this chapter is to present a state-of-the-art survey of the main parallelization ideas and strategies, and to discuss general design principles applicable to all meta-heuristic classes. To achieve this goal, the authors explain various paradigms related to parallel meta-heuristic development, where communications, synchronization, and control aspects are the most relevant. They also discuss implementation issues pointing out the characteristics of shared and distributed memory multiprocessors as target architectures. All these topics are illustrated by the examples from recent literature related to the parallelization of various meta-heuristic methods. Here, the authors focus on Variable Neighborhood Search and Bee Colony Optimization.

Análise socioeconômica de alternativas para o transporte ferroviário de passageiros em uma rede regional

“…A programação paralela consiste na atribuição de tarefas a múltiplas máquinas simultaneamente para resoluçaõ de um problema comum, e tem sido utilizada recorrentemente para resoluçaõ de problemas de otimização (Crainic, 2012). O processamento de dados é realizado em um conglomerado de computadores de alto desempenho interligados em uma rede acessada pela Internet, cuja oferta de serviços con igura o termo "Computaçaõ em Nuvem".…”

Bee Colony Optimization - part I: The algorithm overview

“…Parallelization on all levels for various bees' algorithms can be found in recent literature [4,10,12,35,38,45]. Here we briefly recall the parallelization strategies for BCO proposed in [10,12], and systematically reviewed in [9].…”