Evolutionary deployment optimization for service-oriented software in cloud}{Evolutionary deployment optimization for service-oriented software in cloud

This paper proposes a new deployment optimization method, to address the drawbacks of existing deployment optimization methods, for optimizing the deployment architectures of service-oriented software in cloud. Examples of these drawbacks are the lack of scalability of service instances and virtual machine instances, and the inability to guarantee the solving quality. The method first constructs a deployment optimization model with the goals of improving the running performance and reducing the operation cost of service-oriented software. Next, a genetic-based algorithm MGA-DO is utilized for solving the model. The MGA-DO adopts a group-based encoding scheme to encode the deployment architectures of service-oriented software and combines this scheme with a group-based crossover operator to realize the scalability of service instances and virtual machine instances in the optimization process. Moreover, the MGA-DO utilizes the existing knowledge of deployment optimization to design five types of local search rules, to further improve the local search ability of the algorithm and accelerate the convergence speed. Finally, a series of simulations show that, compared with the existing algorithms, the MGA-DO algorithm performs better on solving the research problem.