How a Network of Processors can Schedule its Work

Abstract

The problem addressed in this paper is to design a method by which a network of processors confronted with a flow of tasks may distribute the computing to be done among the processors so as to make effective use of them to perform the required computations. The method, and its variants, presented in this paper gives weight to the objectives of carrying out the prescribed tasks in short time, and to the relative urgencies associated with those tasks. This problem is reminiscent of the problem of scheduling the flow of jobs through a machine shop. The methods presented here are adapted from a method developed for that problem which were described in [1].