Analyzing synchronous and asynchronous parallel distributed genetic algorithms
Future Generation Computer Systems - Special issue on bio-impaired solutions to parallel processing problems
SETI@home: an experiment in public-resource computing
Communications of the ACM
Distributed Genetic Algorithms
Proceedings of the 3rd International Conference on Genetic Algorithms
Proceedings of the 5th International Conference on Genetic Algorithms
PPSN III Proceedings of the International Conference on Evolutionary Computation. The Third Conference on Parallel Problem Solving from Nature: Parallel Problem Solving from Nature
Distributed and Persistent Evolutionary Algorithms: A Design Pattern
Proceedings of the Second European Workshop on Genetic Programming
Browser-based distributed evolutionary computation: performance and scaling behavior
Proceedings of the 9th annual conference companion on Genetic and evolutionary computation
MapReduce: simplified data processing on large clusters
Communications of the ACM - 50th anniversary issue: 1958 - 2008
Soft Computing - A Fusion of Foundations, Methodologies and Applications - Special issue (1143 - 1198) " Distributed Bioinspired Algorithms"; Guest editors: F. Fernández de Vega, E. Cantú-Paz
Data-intensive computing for competent genetic algorithms: a pilot study using meandre
Proceedings of the 11th Annual conference on Genetic and evolutionary computation
A distributed pool architecture for genetic algorithms
CEC'09 Proceedings of the Eleventh conference on Congress on Evolutionary Computation
A-Teams and Their Applications
ICCCI '09 Proceedings of the 1st International Conference on Computational Collective Intelligence. Semantic Web, Social Networks and Multiagent Systems
Resilience to churn of a peer-to-peer evolutionary algorithm
International Journal of High Performance Systems Architecture
CouchDB: The Definitive Guide Time to Relax
CouchDB: The Definitive Guide Time to Relax
Implementation matters: programming best practices for evolutionary algorithms
IWANN'11 Proceedings of the 11th international conference on Artificial neural networks conference on Advances in computational intelligence - Volume Part II
Pool-Based distributed evolutionary algorithms using an object database
EvoApplications'12 Proceedings of the 2012t European conference on Applications of Evolutionary Computation
Hi-index | 0.00 |
This paper focuses on the iterative design of SofEA , an architecture for distributing evolutionary algorithms (EAs) across computer networks in an asynchronous and decentralized way. SofEA is based on a pool architecture implemented on an object store, allowing the asynchronous interaction with which several clients. The fact that each client is autonomous leads to complex behavior, which will be examined in the work, so that the design can be validated, rules of thumb can be extracted, and the limits of scalability can be found. In this paper we advance the design of an asynchronous, fault-tolerant, and hopefully scalable distributed EA based on the object store CouchDB. We do so by iteratively analyzing running time and average evaluations to solutions on increasingly better versions of the algorithm, looking for the best results, at least from the point of view of running time. By doing so, we increase speed almost fourfold, and also decrement the average number of evaluations to solution in some cases. Experiments have shown also which critical parameters have the bigger influence on the performance in this kind of systems: live population size and number of conflicts, with both being influenced by the number of clients and the size of the population block each client handles at a time. These experiments also show that there is a balance between scalability and fault tolerance, with scalability dropping when a certain number of clients is reached; further clients only increase fault tolerance, at least in the configurations we are using in this paper. The paper also shows that experimentation and measurement conform a good methodology for the design of this kind of asynchronous, heterogeneous and distributed systems, where analytic performance prediction is almost impossible.