A mobile agent based approach for optimizing real-time geographically distributed control systems using swarm intelligence
Letha Hughes Etzkorn, Craig Eichelkraut · 2008
Many geographically distributed control systems have an optimization requirement that is computationally complex and therefore cannot be computed within a deterministic schedule. A real-time distributed control system can have a partial system failure, and during the failure the control system must continue to function in some degraded fashion. To solve these distributed problems, most control systems use either a centralized global algorithm that can be affected by partial failures, or a local greedy algorithm that is robust but non-optimal. Mobile agent algorithms are another means of implementing a distributed approximation. Since mobile agents respond well to dynamic environments and tend to recover quickly from unexpected conditions, they appear well suited to distributed control systems. In this research, we define an approach that employs a Mobile Agent algorithm for optimization real-time systems. We also create an Architecture Description Language to make mobile agents easier to use in distributed control systems. The research results provide a compelling case that mobile agent-based systems have significant merit for optimizing globally distributed real-time systems.