Process scheduling using genetic algorithms
Pai-Chou Wang, Willard Korfhage · 2002
This paper presents a genetic algorithm using a matrix genome encoding to schedule distributed tasks, represented by a directed acyclic graph, on processors in order to minimize the maximum task finishing time. Our experimental results show that this algorithm provides better scheduling results than list scheduling with insertion; and dominant sequence clustering heuristics. Our algorithm generates good schedules even in those cases when the heuristically-generated schedules are worse than using a single processor.