Reinforcement Learning For Scheduling
                    In Parallel Computing Systems 
 
                          Albert Y. Zomaya
                  The University of Western Australia
                                and
                    The University of Missouri-Rolla 
 
 
Abstract
________
 
 
Task scheduling is essential for the proper functioning of parallel 
processor systems. Static scheduling of tasks onto networks of parallel 
processors is well-defined and documented in the literature. However, in 
many practical situations a priori information about the tasks that need to 
be scheduled is not available. In such cases, tasks usually arrive dynamically 
and the scheduling process should be performed "on-the-fly." In this talk, 
I will present a framework based on stochastic reinforcement learning which 
is usually usedd to solve optimization problems in a simple and effective way. 
The use of reinforcement learning reduces the dynamic scheduling problem 
to that of learning a stochastic approximation of an unknown average 
error surface.
 
The main advantage of the proposed approach is that no prior information is 
required about the parallel system under consideration. The learning system 
develops an association between the best action (schedule) and the current 
state of the environment (parallel system). The performance of reinforcement 
learning will be demonstrated by a case study.
 


Biography
_________
 
Albert Y. Zomaya is a Professor in the department of electrical 
and electronic engineering at the University of Western Australia, 
where he also leads the Parallel Computing Research Laboratory. He 
received his PhD from Sheffield University, U.K. The author/co-author of more 
than 80 publications in technical journals, collaborative books, 
and conferences. He is also the author of two books and the editor of two 
volumes. Dr. Zomaya is an associate editor for the IEEE Transactions on 
Systems, Man, and Cybernetics, the International Journal in Computer 
Simulation,  the Journal of Parallel Algorithms and Applications, 
the International Journal on Parallel and Distributed Systems and Networks, and 
the International Journal of Computational Science. He is also the editor 
for the newly-established Wiley Book Series on Parallel and Distributed 
Computing.