Scheduling with Deadlines and Loss Functions

Abstract

The problem of this paper is that of scheduling several one-stage tasks on several processors, which are capable of handling the tasks with varying degrees of efficiency, to minimize the total loss, which is a sum of losses associated with the individual tasks. Each task has a deadline; the individual loss associated with it is a function of amount of time between the deadline and the time of completion if the former precedes the latter, and zero otherwise.