To search, Click below search items.


All Published Papers Search Service


Q-Learning applied to the problem of scheduling on heterogeneous architectures


Younes Hajoui, Omar Bouattane, Mohamed Youssfi and Elhocein Illoussamen


Vol. 18  No. 2  pp. 153-159


Grid computing exploits linked heterogeneous resources to deal with greedy applications or complicated executing tasks. Hence, an efficient resource allocation algorithm is extremely important for the sake of reducing and minimizing the overall execution time. Our approach consists in developing a novel collaborative Q-Learning scheduler using a Holonic Multi Agent System to solve the job scheduling problems on heterogeneous distributed systems. The results show that for many tested dynamic environments, the proposed load balancer optimizes the distribution of tasks and reduces the total tardiness.


Job scheduling, Grid Computing, Distributed System, Multi-Agent system, Load balancer, Q-learning.