Foundations and Architectures for Agent Systems

The proposal is to study the theoretical foundations of, and architectures for, agent systems. In particular, modal higher-order logic as a formalism for knowledge representation and reasoning, and machine learning techniques to handle adaptivity in agents will be investigated. This research is significant because it will provide underpinnings for the new generation of software systems based on agent-oriented concepts that will be considerably more sophisticated than most existing software systems. The expected outcomes are a unified view of probability, logic and learning that will provide theoretical foundations for agent systems of the future and novel architectures for these systems.