The job assignment problem: A study in parallel and distributed machine learning

Abstract

This article describes a parallel and distributed machine learning approach to a basic variant of the job assignment problem. The approach is in the line of the multiagent learning paradigm as investigated in distributed artificial intelligence. The job assignment problem requires to solve the task of assigning a given set of jobs to a given set of executing nodes in such a way that the overall execution time is reduced, where the individual jobs may depend on each other and the individual nodes may differ from each other in their execution abilities. Experimental results are presented that illustrate this approach.