On Sequential Dispatching Policies

Abstract

Job dispatching appears in many communication and computer systems. In this paper, we present a classification of dispatching policies according to their scalability, and introduce the class of sequential dispatching policies. These, highly scalable policies are a special case of index policies. By means of numerical examples, we show that efficient, practically optimal, dispatching may be achieved by a sequential policy, or a simple dynamic variant of such a policy, under heavy load. These numerical examples suggests that efficient and scalable load balancing can often be realized by simple primitive dispatching elements amenable also for machine learning based approaches.