Paralex: an environment for parallel programming in distributed systems
ICS '92 Proceedings of the 6th international conference on Supercomputing
Practical parallel algorithms for dynamic data redistribution, median finding, and selection (preliminary draft)
Isosurfacing in span space with utmost efficiency (ISSUE)
Proceedings of the 7th conference on Visualization '96
The MOSIX multicomputer operating system for high performance cluster computing
Future Generation Computer Systems - Special issue on HPCN '97
The grid: blueprint for a new computing infrastructure
The grid: blueprint for a new computing infrastructure
Design and Implementation of DDH: A Distributed Dynamic Hashing Algorithm
FODO '93 Proceedings of the 4th International Conference on Foundations of Data Organization and Algorithms
Chained Declustering: A New Availability Strategy for Multiprocessor Database Machines
Proceedings of the Sixth International Conference on Data Engineering
Self-organization through bottom-up coalition formation
AAMAS '03 Proceedings of the second international joint conference on Autonomous agents and multiagent systems
Knowledge and Information Systems
Adaptive load balancing: a study in multi-agent learning
Journal of Artificial Intelligence Research
Coalition formation among bounded rational agents
IJCAI'95 Proceedings of the 14th international joint conference on Artificial intelligence - Volume 1
Speciation as automatic categorical modularization
IEEE Transactions on Evolutionary Computation
Hi-index | 0.01 |
We study coalition formation for adaptive dynamic processor load-balancing in distributed systems. The compute nodes are rational agents that are constantly trying to minimize the difference of their CPU loads from the average load of nodes in the vicinity. Each agent uses a game-like evolutionary process to adjust its load by distributing jobs to neighboring nodes and by joining or forming a coalition. Since agents only communicate with their neighbors, the communication cost involved in sharing the load-information with others is greatly reduced while optimal load-balancing is achieved.