基于BPSO-IRS算法的智能值守机器人资源调度方法研究

Research on Resource Scheduling Methods for Intelligent Duty-robot Based on BPSO-IRS Algorithm

针对电网区域内多配电站的巡检,单一值守机器人无法满足区域协调能力且会造成众多资源的浪费,云机器人可以实现资源共享提升巡检效率,以优化系统负载均衡、各类资源利用率均衡和系统能耗为目标,提出了云架构的资源调度系统。搭建资源调度数学模型,结合物理机的几种重要参数建立面向不同目标的数学模型。创建两种不同类型的RVM服务虚拟机:初始放置与增量放置。在云端初始创建服务虚拟机时产生的资源分配问题,可以借鉴PSO算法采用BPSO算法进行放置;在工作环境中对于虚拟机数量不足需要增加虚拟机个数时的分配问题,采用IRS算法进行分配。最后利用设计仿真实验,对比OpenStack原生的LFF调度算法,对BPSO-IRS算法优越性进行验证。

For the patrol inspection of multiple distribution stations in the power grid area,a single duty robot cannot meet the regional coordination ability and cause waste of many resources,so cloud robotics can achieve resource sharing to improve patrol efficiency.Therefore,a resource scheduling system is proposed based on cloud architecture with the goal of optimizing system load balance,resource utilization balance and system energy consumption.Firstly,a mathematical model is established for resource scheduling,and several important parameters of the physical machine combined to establish mathematical models for different objectives.Secondly,two different types of RVM service virtual machines are created:initial placement and incremental placement.The resource allocation problem generated during the initial creation of service virtual machines in the cloud can be placed using the PSO algorithm and BPSO algorithm for reference;when there is a shortage of virtual machines in the work environment and it is necessary to increase the number of virtual machines,the IRS algorithm is used for allocation.Finally,using design simulation experiments,the superiority of the BPSO-IRS algorithm is verified by comparing it with the native LFF scheduling algorithm of OpenStack.