Concurrent differential evolution based on generational model for multi-core CPUs

Authors:
Kiyoharu Tagawa
Affiliations:
School of Science and Engineering, Kinki University, Higashi, Osaka, Japan
Venue:
SEAL'12 Proceedings of the 9th international conference on Simulated Evolution and Learning
Year:
2012

Citing 9
Cited 1

Efficient and Accurate Parallel Genetic Algorithms

Efficient and Accurate Parallel Genetic Algorithms
Differential Evolution – A Simple and Efficient Heuristic for Global Optimization over Continuous Spaces

Journal of Global Optimization
Java Concurrency in Practice

Java Concurrency in Practice
Differential Evolution: In Search of Solutions (Springer Optimization and Its Applications)

Differential Evolution: In Search of Solutions (Springer Optimization and Its Applications)
The Art of Concurrency: A Thread Monkey's Guide to Writing Parallel Applications

The Art of Concurrency: A Thread Monkey's Guide to Writing Parallel Applications
Fast parallelization of differential evolution algorithm using MapReduce

Proceedings of the 12th annual conference on Genetic and evolutionary computation
Parallelism and evolutionary algorithms

IEEE Transactions on Evolutionary Computation
Differential Evolution: A Survey of the State-of-the-Art

IEEE Transactions on Evolutionary Computation
Improving Classical and Decentralized Differential Evolution With New Mutation Operator and Population Topologies

IEEE Transactions on Evolutionary Computation

Adapting evolutionary algorithms to the concurrent functional language Erlang

Proceedings of the 15th annual conference companion on Genetic and evolutionary computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

In order to utilize multi-core CPUs more effectively, a new Concurrent Differential Evolution (CDE) is proposed. Then the proposed CDE (CDE/G) is compared with a conventional CDE (CDE/S). CDE/S uses only one population because it is based on the steady-state model. Therefore, CDE/S requires a time-consuming mutual exclusion or "lock" for every read-write access to the population. On the other hand, CDE/G is based on the generational model. By using a secondary population in addition to a primary one, CDE/G does not require any lock on the population and therefore is faster. Through the numerical experiment and the statistical test, it is demonstrated that CDE/G is superior to CDE/S in not only the run-time but also the quality of solutions.