面向人机物融合的泛在计算正成为软件发展的新需求和新趋势, 基于这种新形态计算模式的人机物融合应用将软件技术进一步拓展至对线下资源, 包括物理设备和人力资源的有效利用. 作为典型的人机物融合场景, 物理世界中设备资源与人力资源间的协作具有资源可选性、任务高频性、工人动态性的特点, 传统的资源调度技术无法有效应对该类型任务(简称为DHRC任务)中的调度需求. 为此, 提出一种面向设备与人力资源协作任务的优化调度方法, 所提方法分为设备资源调度和人力资源调度两个阶段. 在设备资源调度阶段, 提出基于NSGA-II 的设备资源调度算法, 在综合考虑任务距离、设备负载和设备位置周边工人人数等因素情况下实现任务对资源的优化选择. 在人力资源调度阶段, 提出基于DPSO的人力资源调度算法, 根据工人位置和协作依赖等因素实现对工人的优化选择以及相应的路径规划. 在模拟环境内的实验结果表明, 所提方法第1阶段的算法在效率上与对比算法相当, 在效用性上优于对比算法(离散粒子群优化算法). 第2阶段的算法在效率上与效用性上均优于对比算法(使用锦标赛机制改进的遗传算法).
Ubiquitous computing for human-cyber-physical integration is becoming a new requirement and trend in software development. Based on this new computing paradigm, human-cyber-physical applications further extend software technology to the effective utilization of offline resources, including physical devices and human resources. As a typical human-cyber-physical scenario, the collaboration between the device and human resources in the physical world features resource selectivity, high task frequency, and worker dynamics. Traditional resource scheduling techniques cannot meet the scheduling requirements of this task type (referred to as DHRC task). Thus, this study proposes an optimal scheduling method for collaborative tasks between device and human resources. This method includes two stages of device resource scheduling and human resource scheduling. In the device resource scheduling stage, a device resource scheduling algorithm based on NSGA-II is proposed to optimize task resource selection by comprehensively considering such factors as task distance, device load, and the worker number around the device location. In the human resource scheduling stage, a human resource scheduling algorithm based on DPSO is put forward to optimize the worker selection and corresponding path planning according to such factors as worker location and collaboration dependency. Experiments in a simulated environment show that the algorithm in the first stage is equivalent in efficiency and superior in utility to the compared algorithm (discrete particle swarm optimization algorithm). The algorithm in the second stage is superior in efficiency and utility to the compared algorithm (the genetic algorithm improved by the tournament mechanism).