Decentralized Control Modeling by Feasible Decomposition Method

Application of the traditional centralized approach to construction of automated optimal control systems for complex industrial facilities is often difficult due to the high dimensionality and complexity of the emerging control problems. Such industrial facilities, in particular, can include technological processes of finished products and production complexes of the petrochemical and oil and gas industry. One of the effective ways to eliminate these difficulties can be the use of decentralized control systems with a multi-level hierarchical structural organization. The mathematical basis of decentralized hierarchical control is formed by the methods of optimal control problems decomposition. One of the most effective and promising method for practical use in this class is the feasible decomposition method. This method is distinctive for its obvious operating principle and visual physical interpretation. However, it is developed only at the conceptual level, and it is difficult to apply it in practice, since there are no specific implementation mechanisms, in particular, algorithms for working interaction between the coordinating body and local control systems. In the paper a possible approach to the elimination of this problem is presented. The findings of the experimental study confirm the validity and effectiveness of this approach.