Using evolutionary programming to schedule tasks on a suite of heterogeneous computers

A virtual heterogeneous machine can be used to process tasks more quickly than a suite of homogeneous machines because jobs can be scheduled on machines that are particularly effective for each job’s computational requirements. Such scheduling presents a combinatorial optimization problem. This paper assesses various greedy algorithm and evolutionary programming approaches over sets of randomly generated scheduling problems. The results indicate the potential for effectively using a hybrid scheduling method that takes advantage of both greedy and evolutionary optimization.