In High Performance Computing, every processor is a player in a symphony of speed—and harmony matters as much as power. This research theme explores how we can intelligently distribute computational tasks across clusters and parallel architectures so that no processor idles while others overflow. We design, implement, and evaluate dynamic load balancing and scheduling algorithms that learn, adapt, and optimize in real time. Through reinforcement learning, multi-agent systems, and custom scheduling frameworks, we aim to transform computation from a static sequence of operations into an evolving collaboration among machines.

Our work in this area pushes beyond traditional heuristics. It envisions load balancing as a form of collective intelligence, where each computing node not only performs its task but also participates in a decision-making process that improves global performance. In doing so, we move closer to self-organizing, autonomic systems that embody the spirit of adaptive parallelism.

1. Dynamic Load Balancing with Learning Agents (Invited Paper)
   J.P. Pabico
   Proceedings of 2009 Europe-Philippines International Workshop on Modeling, Simulation, and Grid Computing (MODEL 2009)
   Holy Name University, Tagbilaran City, Bohol, 04-06 May 2009 

2. Dynamic Load Balancing Algorithms for Embarrassingly Parallel Tasks (Invited Paper)
   J.P. Pabico
   H.N. Adorna and R.P. Saldaña (eds.) Proceedings of the 9th Philippine Computing Science Congress (2009 PCSC), pp. 70
   Silliman University, Dumaguete City, 2-3 March 2009
   PDF 

3. A Framework for Multiagent-Based Scheduling of Parallel Jobs
   J.P. Pabico
   R.P. Saldaña (ed.) Proceedings of the 6th Philippine Computing Science Congress (PCSC 2006) p. 81-88
   Ateneo De Manila University, Loyola Heights, Quezon City, 28-29 March 2006
   arXiv:1506.07964v1 [cs.DC] | PDF 

4. Automatic Selection of Loop Scheduling Algorithms Using Reinforcement Learning
   S. Dhandayuthapani, I. Banicescu, R.L. Cariño, E. Hansen, J.P. Pabico, and M. Rashid
   Proceedings of the International Workshop on Challenges of Large Applications in Distributed Environments (HPDC-CLADE 2005) pp. 87-94
   Research Triangle Park, North Carolina, USA, 24 July 2005 

5. Performance Evaluation of a Dynamic Load Balancing Library for Cluster Computing
   M. Balasubramaniam, R.L. Cariño, I. Banicescu, and J.P. Pabico
   International Journal of Computational Science and Engineering, vol 1, nums 2/3/4 (Special Issue on Adaptivity in Parallel Scientific Computing), pp 118-133, May 2005
   (ISSN 1742-7185, DOI 10.1504/IJCSE.2005.009697) 

6. Design and Implementation of a Novel Dynamic Load Balancing Library
   I. Banicescu, R.L. Cariño, J.P. Pabico and M. Balasubramaniam
   Parallel Computing, vol 31, num 7 (special issue on Heterogeneous Computing), pp 736-756, July 2005
   (ISSN 0167-8191, DOI 10.1016/j.parco.2005.04.006) 

7. Overhead Analysis of a Dynamic Load Balancing Library for Cluster Computing
   R.L. Cariño, M. Balasubramaniam, I. Banicescu, and J.P. Pabico
   Proceedings (CDROM) of the 19th International Parallel and Distributed Processing Symposium (IPDPS2005), 14th International Heterogeneous Computing Workshop (IPDPS-HCW 2005) p. 122b
   Denver, Colorado, USA, 4-8 April 2005