Fully automatic manufacturing systems have received wide attention and are being adopted rapidly in manufacturing industry with the application of Internet of Things. How to coordinate correlated operations efficiently has become a prominent challenge for many practitioners. In this regard, this paper considers an optimal control issue arising in a fully automatic manufacturing system based on Internet of Things and formulates it as a novel parallel machine scheduling problem with multiple servers. The setup stage and the removal stage of a job must be performed by one of the servers on one of the machines while the processing stage of the job can be performed automatically by the machine. The problem takes maximum slack times, machine availability, server availability, machine eligibility, and server eligibility into consideration simultaneously and its objective is to minimize makespan. The problem generalizes a number of problems in the topic of server scheduling. To obtain optimal solutions, a mixed integer programming model is provided. To facilitate the development of heuristic algorithms, the classical list-and-assign technique in this new problem setting is deeply analyzed. Six heuristic algorithms are proposed based on the technique. Numerical experiment shows that it is advantageous to list the jobs according to non-increasing order of setup time or according to non-increasing order of the sum of setup time and removal time.

全自动制造系统已受到广泛关注，并随着物联网的应用在制造业中迅速采用。如何有效地协调相关操作已成为许多从业人员面临的重大挑战。在这方面，本文考虑了基于物联网的全自动制造系统中出现的最优控制问题，并将其表述为具有多个服务器的新型并行机器调度问题。作业的设置阶段和删除阶段必须由一台机器上的服务器之一执行，而作业的处理阶段可以由机器自动执行。该问题需要最长的松弛时间，计算机可用性，服务器可用性，计算机资格，和服务器资格同时考虑，其目标是最大程度地缩短制造时间。该问题概括了服务器调度主题中的许多问题。为了获得最佳解，提供了一个混合整数规划模型。为了促进启发式算法的发展，对这种新问题中的经典列表分配技术进行了深入分析。基于该技术提出了六种启发式算法。数值实验表明，按照设置时间的不递增顺序或按照设置时间和清除时间之和的不递增顺序列出作业是有利的。为了促进启发式算法的发展，对这种新问题中的经典列表分配技术进行了深入分析。基于该技术提出了六种启发式算法。数值实验表明，按照设置时间的不递增顺序或按照设置时间和清除时间之和的不递增顺序列出作业是有利的。为了促进启发式算法的发展，对这种新问题中的经典列表分配技术进行了深入分析。基于该技术提出了六种启发式算法。数值实验表明，按照设置时间的不递增顺序或按照设置时间和清除时间之和的不递增顺序列出作业是有利的。