Online scheduling: a survey

Abstract

In this article a deep search of the literature of online scheduling is conducted. This paper intends to assess the developments and solutions found for online scheduling problems. Online scheduling is a very important topic since most of the real scheduling problems have dynamic characteristics. First, it was developed a literature review about scheduling problems, dividing them in stochastic and deterministic problems as well as in online and offline problems. Then, a bibliometric analysis was performed. Finally, some case studies in the field of online scheduling were analyzed. Online Scheduling is mostly explored in industry and health areas. In some articles explored there is a rescheduling, and the sequence of task may change due to the arrival of new tasks. In other cases, the new tasks are introduced in blocks of time that do not affect the previous schedule. This last technique is limited, since, with the arrival of new tasks, the schedule is not re-evaluated. Therefore, it is thought that, in future work, within the scope of online scheduling, when new tasks or other significant changes enter the system, the system should be evaluated, allowing the necessary changes to be made to the existing schedule. The Industry 4.0 and the evolution of Internet of Things (IoT), Deep Learning and Machine Learning favours a continuous and real-time flow of information, which allows the implementation of real-time online scheduling. This is a branch that should be explored in future works.