A novel resource aware scheduling with multi-criteria for heterogeneous computing systems

Multi-criteria scheduling for heterogeneous computing systems (HCSs) is one of the prime concerns of modern computing world. It is an NP-complete problem. In this paper, we have proposed two resource aware multi-criteria scheduling, where scheduling is done by considering real-time resources such as, clock frequency, memory capacity and residual energy. The proposed algorithms are designed using four conflicting objectives, viz., minimization of makespan, energy-consumption, load-balancing, and maximization of resource utilization. The proposed work is analyzed and validated by extensive simulations with synthetic as well as benchmark data sets. Through the simulation results, it is observed that the proposed work has considerable improvements over Min-Min, MCT, Genetic algorithm (GA) and priority based performance improved algorithm (PPIA). A statistical hypothesis test Analysis of Variance (ANOVA) and post hoc test are performed to demonstrate the effectiveness of the proposed work. (C) 2018 Karabuk University. Publishing services by Elsevier B.V.