Title: A generalised control metaheuristic framework for industrial processes

Authors: F.N. Koumboulis, M.P. Tzamtzi

Addresses: Department of Automation, Halkis Institute of Technology, Psahna, Evia 34400, Greece. ' Department of Automation, Halkis Institute of Technology, Psahna, Evia 34400, Greece

Abstract: A metaheuristic control design scheme is proposed to solve control design problems formulated as optimisation under constraints. Actuator, state and/or output variable constraints are considered. The algorithm utilises process simulation blocks without specific characteristics. Candidate controllers are randomly selected within a search area, whose range and location change adaptively. The proposed design framework may be applicable to many industrial applications, since it does not use any assumptions regarding the structure of the controlled plant or the controller or the performance indices. Application of the framework requires the following main steps: (1) determination of a cost criterion, whose minimisation corresponds to desired closed-loop performance; (2) formulation of any constraints on the plant|s variables; (3) determination of any other desired and algorithmically checkable closed-loop characteristics; (4) development of closed-loop simulation blocks, based on the non-linear plant|s model or its linear approximants and (5) determination of the initial search area of the controller parameters.

Keywords: industrial automation; intelligent control; optimisation methods; nonlinear systems; search methods; industrial processes; metaheuristics; control design; simulation.

DOI: 10.1504/IJAAC.2010.030812

International Journal of Automation and Control, 2010 Vol.4 No.2, pp.218 - 233

Published online: 06 Jan 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article