We investigate a multi-skill technician and reassignable task scheduling problem in a cloud computing company. In the problem, multi-skill technicians are assigned to process a large number of tasks from customer requests in a certain scheduling horizon. The tasks are allowed to be reassigned to another technician multiple times, and one technician can process multiple tasks in parallel. The company not only focuses on processing efficiency, but also expects to improve customers’ experience and technicians’ satisfaction. We characterize the feasible solutions and introduce a weighted objective with three metrics: processing efficiency, response delay, and workload balance. An effective two-stage hierarchical optimization method embedded in a greedy randomized adaptive search procedure framework is proposed. In the first stage, initial solutions are generated by a greedy randomized construction procedure, and then improved by local search with an ejection chain operator to optimize processing efficiency. In the second stage, two local search procedures with five operators for improving response delay or workload balance are designed. Computational experiments are conducted to evaluate the effectiveness of our algorithm. The results show that the proposed algorithm is competent in fast computing a schedule of high quality. It also reveals that reassignments are helpful in reducing response delay and balancing workloads in the scheduling.

中文翻译：

---

在云计算公司中安排多技能技术人员和可重新分配的任务


我们调查了一家云计算公司的多技能技术人员和可重新分配的任务调度问题。在该问题中，分配了多技能技术人员在一定的调度范围内处理来自客户请求的大量任务。允许将任务多次重新分配给其他技术员，并且一个技术员可以并行处理多个任务。该公司不仅注重加工效率，还希望提高客户体验和技术人员的满意度。我们描述了可行的解决方案，并引入了一个加权目标，其中包含三个指标：处理效率、响应延迟和工作负载平衡。该文提出一种嵌入贪婪随机自适应搜索程序框架中的有效两阶段分层优化方法。在第一阶段，通过贪婪的随机构造过程生成初始解决方案，然后通过使用顶出链运算符进行本地搜索进行改进，以优化处理效率。在第二阶段，设计了两个本地搜索程序，具有五个运算符，用于改善响应延迟或工作负载平衡。进行计算实验以评估我们的算法的有效性。结果表明，所提算法能够快速计算出高质量的进度表。它还揭示了重新分配有助于减少响应延迟和平衡计划中的工作量。