Intelligent Agents Based Non -- Square Plants Control

Abstract

Generalized approach for non-square plants (NSP) control based on multi agent system (MAS) is proposed. The control system provides relevant reconfiguration and adaptation in case of plant changes, due to manipulated variables (MV) saturation, filature of input and/or output instrumentation, close proximity to ill-conditioning. A big flexibility is reached when objective function, operation properties and constraints are dynamically changed. Simulation results are presented from MAS realized on JADE platform for two inputs two outputs (TITO) plant using stigmergy as an optimizing method. Comparative analysis of synchronous and asynchronous MAS is carried out