Dynamic task scheduling in distributed real time systems using fuzzy rules

Abstract This paper addresses scheduling on distributed real time systems. A distributed scheduling algorithm is presented. It consists of a guarantee routine, bidder algorithm and a decision maker. The guarantee routine checks whether a newly arrived task is schedulable or not. In order to make it suitable for distributed allocation and scheduling, an optimal guarantee routine is analyzed and extended. A node that cannot guarantee a task, asks for bidding and a decision algorithm evaluates the offers. Bidders send incomplete information about their states and, to handle this, the decision algorithm makes use of fuzzy rules to dispatch the unguaranteed task.