移动云环境面向多重服务选择的计算卸载算法 postprint

Abstract: Mobile cloud computing can use the computation offloading to improve the energy efficiency of mobile devices and the execution delay of applications. However, in the face of the multiple service selection from cloud, the computation offloading decision is a NP problem. In order to address this problem, a genetic algorithm is designed to find the best application partitioning decision solution for computation offloading. In genetic population initialization, our algorithm combines of predefined and random chromosomes to intialize the population, which can reduce the generation of the inefficient chromosomes. At the same time, the algorithm designs a specific fitness function based on Hamming distance function for the predefined reserved populution, which can better measure the difference between chromosomes. The population crossover uses respectively the inbreeding and crossbreeding to enrich individual species. Our algorithm uses the modified genetic operations to reduce the ineffective solutions and obtain the best feasible solution in a reasonable time. We evaluated the efficiency of the proposed algorithm using graphs of real mobile applications in simulation experiments. The evaluated conclusions denote that our designed algorithm has a better performance than the comparision algorithms on the application execution energy, the execution time and the overall weight cost.