Problem independent parallel genetic algorithm for design optimization

Abstract

Genetic Algorithms, originally inspired by the nature, have come a long a way in the past decade to establish itself as a numerical optimization tool of choice. The simplicity of its structure, robustness of its operation, global view of its search technique, efficiency of its sorting though the unknown, all contribute to the excellence of genetic algorithm as an engineering design optimization tool. With the advent of modern parallel and distributed supercomputer systems, there is a renewed enthusiasm about fast genetic algorithms that efficiently solves very large numerical optimization problems utilizing these machines. The working of genetic algorithms is analyzed to extract parallelism in its functionality. The effects of various parallelization proposals and details of a parallel implementation are provided. Several representative problem solutions are presented to demonstrate the utility of the implementation. These indicate a strong future for parallel genetic algorithms for solving a large variety of scientific, engineering, and design problems.

An Investigation of Using Parallel Genetic Algorithm for Solving the Shortest Path Routing Problem

“…Parallel GA has been successfully used in various problems such as design optimization (Atiqullah, 2002) , transport route planning (Meghanathan and Skelton, 2007), time series forecasting (Eklund, 2003;Ourdighi and Benyettou, 2010), network design (Huang et al, 1997) and sorting (Han, 1999).…”

The Analysis and Research of Parallel Genetic Algorithm

“…But the parallel genetic algorithm (PGA) can combine the high speed of parallel computers parallelism with the GA inherent parallelism, which accelerates the search process of GA, maintains the diversity of population prevents premature, and solves such a complicated problem efficiently and effectively. Therefore, the research and application of the PGA have received much attention in recent years [2][3][4]. As a result of the PGA superiority, it will obtain the widespread application and the development in many domains.…”