Adaptive Computing System for Distributed Process Control

Abstract

This paper is dedicated to solve some issues related to climate change that negatively affect the productivity and quality of agricultural products. Agricultural production processes are spatially distributed processes and require a special approach in the development of control systems. The paper proposes the development of an adaptive computing system based on Intelligent Agents. Two types of Agents are defined with functions of perceiving the controlled process and with functions of action on the process. The set of Agents forms a mesh network that ensures the communication between them in order to exchange the decisions taken by each Agent. The decisions generated by the multitude of Agents are based on the application of the knowledge and mathematical models on which the Fuzzy Logic and the Neural Networks are based. The functionality of the Agents and the system as a whole is demonstrated on the basis of the functional schemes and the sequence diagram specifying the mode of communication and the sequence of operations performed by the set of Agents. The adaptation of the system to the agricultural production process takes place by updating the knowledge at each decision-making cycle.