EVOLUTIONARY ALGORITHMS AND THEIR PARALLEL IMPLEMENTATION

EVOLUTIONARY ALGORITHMS AND THEIR PARALLEL IMPLEMENTATION

Abstract

The objective of the present work is to introduce to the various evolutionary algorithms (EAs) and their application in solving combinatorial optimization problems. EAs like Ant Colony System, Genetic Algorithms, Evolutionary Strategies and Evolutionary programming are presented. The use of parallel computing for EAs is also considered.
EAs are generally used to solve optimization problems that are NP-hard. For EAs to give efficient solutions high memory is required to store the large population of solution. Also a high computing power is required for quicker results. Indeed, the use of parallel computers (with dozens of processors) can speed up the execution of EAs and provide the large memory space they require. It is possible to take benefit of the intrinsic parallelism of EAs (e.g., for the concurrent exploration of the search space) in order to design efficient parallel implementations. However each EA has its own characteristics and therefore a general rule cannot be defined.