Design of intelligent component of hierarchical control system

Abstract

One of the main tasks of modern hierarchical systems management are integration of technological, organizational and economic management functions and processes. Another important task is creation of unified information space with accurate, complete and current information. Efficient hierarchical systems management requires intelligent processing of large amounts of heterogeneous information. It is appropriate to process information via intelligent components that are built using artificial neural networks. Strategic information about macroenvironments, microenvironments and internal environments are the input data for intelligent components of upper management levels. Intelligent components may solve different tasks which could have features like: large amount of data, data diversity (quantitative, qualitative, text), contradiction and incomplete data, consistency and high intensity of incoming data, high computing amount with dominance of computing operations over logical operations, recursions and regularity of data processing using neural network algorithms, continuous complications of processing algorithms and increasing requirements for results accuracy, possibility to process data in parallel. The method of synchronized spatio-temporal mapping algorithms for intelligent operation component that provides synchronization of data flow intensity with computing intensity (hardware implementation) and takes into account the processor architecture (software implementation) has been designed. It has been proposed to use following principles: conveyor and spatial parallelism in data processing, modularity, specialization, uniformity and regularity of the structure, programmability architecture. during design of intelligent hardware components. Evaluation of structure of intelligence hardware components carried out using test equipment efficiency. Equipment efficiency takes into account number of interface outputs and number of interneuron connections. At the next step it links performance costs of equipment and evaluates elements by device performance. The method for designing of intelligent component management system that uses synchronized spatio-temporal mapping algorithm has been described. Current method takes into account the components, the requirements of the specific application and provides implementation of intelligent components with high efficiency.