Agent-Based Collaborative Workforce Management
(H. F. Leung)

One of the key properties defining intelligent agent is social ability, which is the capability of an agent to interact with other agents or humans. One way to understand how agents should behave in a particular environment is to model agent interaction in that environment as a game. Game theoretic modelling and analysis are useful to understand how decision-makers interact in different environments. In traditional game theory, outcome and strategy analysis depends on rationality and expected utility. However, in the context of agent interaction, the concept of risk is often applied. Sometimes it is also good for players to making decisions based information observable from past experience, such as opponents’ reputations. It is the aim of this project to introduce a formal definition of the concepts of risk and reputation in game theory, and analyse games and strategies based on these new concepts, trade-off between risk and utility, and the state of risk equilibrium.