EMSS 2014 Proceeding

A multi task assignment method in cloud-based simulation

Authors:   Lei Ren, Hejian Ou, Jin Cui, Bowen Li, Baocun Hou

Abstract

Cloud-based Simulation (Cloud Simulation) can significantly improve the capacity of existing networked modeling and simulation systems. The performance of the Cloud Simulation Platform tightly couples with the partitioning and allocation of simulation modules among different hosts. This paper proposed a Cloud Simulation resources scheduling strategy for initial partitioning based on Genetic Algorithm (GA). The Cloud Simulation resources scheduling model was given by considering the computing resources of involved hosts as well as the predicted computing load of simulation modules. The improved strategy of GA is designed according to the characteristics of the Cloud Simulation multi-task assignment problem. The simulation results demonstrate the feasibility and efficiency of the proposed method.

I3M  Scientific Sponsors

I3M  Industrial Sponsors

I3M  Media Sponsors