An Improved Upper-Bound Algorithm for Non-preemptive Task Scheduling

One of the core problems in real-time systems, finding a feasible schedule for a certain task set, is subject to alternative solutions given the various system constraints. While preemptive scheduling has benefited from a large number of significant results, the non-preemptive case has still room for improvement. This paper extends our previous results regarding the minimum number of processors required for getting a feasible schedule and the possible ways of improving the well known traditional scheduling algorithms in the case of non-preemptive, single-instance, multiprocessor systems. The improvements refer to handling the situations when the already determined minimum number of processors proves insufficient, and also to extending an existing technique which allows the algorithm to overcome certain stalling situations.