An efficient optimization technique for task matching and scheduling in heterogeneous computing systems

摘要

A new optimization technique, the genetic annealing algorithm (GAA), is proposed to solve the task matching and scheduling problem in a heterogeneous computing system. The GAA is simple in design; it employs only the stir operation, a novel idea with the annealing concept, to locate optimal solutions. Experimental evaluation shows that compared with the genetic algorithm, simulated annealing and guided evolutionary simulated annealing approaches, the GAA yields constantly favorable performance in terms of speedup, running time, cost and complexity.