A Model of Context Awareness Agent System Based on Dynamic Fuzzy Logic

One of the main challenges of context-aware computing is how to elicit and represent complex, context-dependent requirements, and use the resulting representation within context-aware applications to support decision-making processes. In this paper, we describe context information with dynamic fuzzy logic and present an Agent-based context-aware system model. Unfortunately, there are often many errors that interfere in the Agent sensors, the process of each Agent communication and the sharing context information of each Agent. In order to robust the context awareness Agent system, we handle these errors with rewards mechanism, so that they do not interfere in the context information learning process. We consider an Agent decider that predicts stock price changes and uses its predictions for selecting actions. The simulation results demonstrate the validity and effectiveness of the proposed algorithm.