ABSTRACT

Volunteer computing systems like BOINC or Comcute are strongly supported by a great number of volunteers who contribute resources of their computers via the Web. So, the high efficiency of such grid system is required, and that is why we have formulated a multi-criterion optimization problem for a volunteer grid system design. In that dilemma, both the cost of the host system and workload of a bottleneck host are minimized. On the other hand, a reliability of this grid structure is maximized. Moreover, genetic programming has been applied to determine the Pareto solutions. Finally, a negative selection procedure to handle constraints has been discussed.

Part 8: Intelligent Distributed SystemsInternational audienceIn this paper, harmony search algorithms have been proposed to self-configuration of intelligent grids for big data processing. Self-configuration of computer grids lies in the fact that new computer nodes are automatically configured by software agents and then integrated into the grid. A base node works due to several configuration parameters that define some aspects of data communications and energy power consumption. We propose some optimization agents that are based on harmony search to find a suboptimal configuration of fault–tolerant grids processing big data. Criteria such as probability that all tasks meet their deadlines and also a reliability of grid are considered. Finally, some experimental results have been considered

No abstract

Abstract. The paper presents our approach to SVM implementation in parallel environment. We describe how classification learning and prediction phases were pararellised. We also propose a method for limiting the number of necessary computations during classifier construction. Our method, named one-vs-near, is an extension of typical one-vs-all approach that is used for binary classifiers to work with multiclass problems. We perform experiments of scalability and quality of the implementation. The results show that the proposed solution allows to scale up SVM that gives reasonable quality results. The proposed one-vs-near method significantly improves effectiveness of the classifier construction.