The Multi-Agent Systems module is designed to provide students with a comprehensive understanding of systems composed of multiple interacting intelligent agents. It aims to equip students with the theoretical foundations and practical skills necessary to design, analyze, and implement multi-agent systems. This module is particularly relevant to students in Information and Computing Science, Computer Science and Technology, and Mechatronics and Robotic Systems, as it bridges concepts from artificial intelligence, distributed systems, and robotics. By exploring topics such as agent architectures, communication protocols, coordination strategies, and decision-making processes, students will be prepared to tackle complex, distributed problems in various domains, including autonomous robotics, distributed control systems, and intelligent software applications.
A. Critique the concept of an agent, comparing and contrasting it with other software paradigms such as object-oriented programming, and analyze the specific characteristics that make certain applications more suitable for an agent-oriented solution. B. Assess the key issues involved in constructing agents capable of intelligent autonomous action, and critically evaluate the main approaches and methodologies used in the development of such agents. C. Evaluate the key considerations in designing societies of agents that can effectively cooperate to resolve problems, including an in- depth analysis of the different types of multi-agent interactions possible in these systems. D. Analyze principal application areas where agent-based solutions are applied, and demonstrate the ability to develop a functional agent-based system using a contemporary agent development platform, showcasing practical proficiency.
The module will be delivered through a combination of lectures, tutorials, and private study to provide a balanced and in-depth learning experience. Lectures will introduce the core theoretical concepts, models, and frameworks underpinning multi-agent systems, guided closely by the textbook by Michael Wooldridge. Tutorials will offer a more interactive environment where students can apply these concepts to practical problems, discuss case studies, and develop small-scale agent-based applications using suitable development tools. These sessions are also intended to encourage collaborative learning and critical thinking. Students will be expected to engage in significant private study to review lecture materials, complete reading assignments, and work on tutorial exercises and project- based tasks, enabling them to deepen their understanding and develop the skills necessary for independent problem-solving in multi-agent system design.