Real-time scheduling with soft deadlines and cost/value functions

The use of a central facility for performing modeling and optimization operations is an approach employed in improving overall performance in manufacturing environments. Such a facility may be thought of to operate using the well established client-server architecmre, wherein the facility performs the role of the server. The prototype system built using die Simplex architecmre [2], defmed by the Software Engineering Institute (SEI) at Carnegie MeUon University and the Hybrid Modellng Facility [3], may be considered to fall under this category.

Faculties such as those mentioned above are essentially real-time systems and are used largely in a manufacturing oriented shop floor environment. Such systems usually behave more as soft real-time systems rather than hard ones. The reason for this, is that, the failure to meet deadlines of tasks in the system may not lead to catastrophic failure.

The scheduler forms an integral part of such systems and essentially decides on the execution ordering of the tasks, corresponding to the requests received. This ordering is dependent, in turn, on the algorithm that is used.

The aim of this study is to improve the operation of a simulated system based on the above mentioned architectures, by providing appropriate scheduling algorithms for the ordering of tasks that are to be executed, in response to requests from clients. By appropriate we mean that the scheduling algorithm will take into account the fact that tasks in the system operate under soft deadlines rather that hard ones, and also incorporate the cost or value factors that may be associated with the requests.