Quasi-static scheduling for multiprocessor real-time systems with hard and soft tasks

We address in this paper the problem of scheduling for multiprocessor real-time systems with hard and soft tasks. Utility functions are associated to soft tasks to capture their relative importance and how the quality of results is affected when a soft deadline is missed. The problem is to find a task execution order that maximizes the total utility and guarantees the hard deadlines. In order to account for actual execution times, we consider time intervals for tasks rather than fixed execution times. A single static schedule computed offline is pessimistic, while a purely online approach, which computes a new schedule every time a task completes, incurs an unacceptable overhead. We propose therefore a quasi-static solution where a number of schedules are computed at design-time, leaving for run-time only the selection of a particular schedule, based on the actual execution times. We propose an exact algorithm as well as heuristics that tackle the time and memory complexity of the problem. We evaluate our approach through synthetic examples and a realistic application.