Periodic Scheduling with Costs Revisited

Abstract

Periodic broadcast scheduling typically considers a set of discrete data items, characterized by their popularity, size and scheduling cost. A classic goal is the definition of an infinite, periodic schedule that yields minimum mean client serving time and minimum mean scheduling cost at the same time. This task has been proven to be NP-Hard and more recent works have discarded the scheduling cost attribute, focusing only on the minimization of the mean client serving time. In the context of the present work the scheduling cost is reinstated. An analysis-based scheduling technique is presented, which can practically minimize the mean client serving time and the mean scheduling cost concurrently. Comparison with related approaches yields superior performance in all test cases.