On-demand bounded broadcast scheduling with tight deadlines

We investigate an online scheduling problem motivated by pull-based data delivery systems where there is a server keeping a number of pages; and clients requesting the same page can be satisfied simultaneously by one broadcast. We focus on the special case where preemption is allowed but aborted requests can never be satisfied again. The HEU algorithm of Woeginger [10] is proven to be optimal in maximizing the number of satisfied requests when the pages have equal length and the requests have tight deadlines. However, we show that when there are maximum bounds on the number and weight of requests at any time in the system, the HEU algorithm is not optimal. We then propose a modified algorithm, VAR, which is optimal for this case.