采用改进二进制蝙蝠算法的任务调度算法

A Task Scheduling Algorithm Based on an Improved Binary Bat Algorithm

针对海服务系统环境中任务完成时间过长的问题,提出一种基于改进二进制蝙蝠算法(IBBA)的任务调度算法。该算法通过在二进制蝙蝠算法(BBA)寻优过程中引入非线性惯性权重因子,平衡全局搜索和局部搜索能力;利用2个不同的邻居蝙蝠算子构成扰动项,避免算法陷入局部最优;利用自适应学习因子调整全局最优算子和邻居蝙蝠算子的权重,控制整个寻优过程由全局搜索为主向局部搜索为主过渡。实验结果表明:IBBA可以稳定地搜索到全局最优值,与现有的基于二进制粒子群算法(BPSO)和二进制蝙蝠算法的任务调度算法相比,在任务数较多时,基于IBBA的任务调度算法能避免过早收敛,使任务完成时间显著减少。在海服务网络中,将该算法应用于任务调度,可提高处理大粒度服务的效率。

A task scheduling algorithm based on an improved binary bat algorithm（IBBA-TA）is proposed to solve the problem of long completion time of tasks in sea service environments.The algorithm introduces nonlinear inertia weight factors in the optimization process of the binary bat algorithm（BBA）to balance capabilities of global and local searches.A perturbation term is constructed by using two mutually exclusive neighbor bats to avoid local optimums.Weights of both the global optimal operator and the neighbor bat operator are adjusted using an adaptive learning factor,and control the transition of the optimization process from global searches to local searches.Experimental results show that IBBA-TA stably obtains the global optimal value.Comparisons with the existing task scheduling algorithms based on the binary particle swarm optimization algorithm（BPSO）and the binary bat algorithm show that when the number of tasks is large,IBBA-TA avoids premature convergence and significantly reduces completion time of tasks.It is concluded that the algorithm can be used for task scheduling to improve the efficiency of processing large granularity services in sea service networks.