Piotr J. Gmytrasiewicz
In our work we adopted the decision-theoretic principle of expected utility maximization as a paradigm for designing autonomous rational agents operating in multi-agent environments. Our approach diifers from techniques based on game theory; we are not looking for equilibria, and we do not have to assume that the agents have arrived at the state of common knowledge. Instead, we endow an agent with a representation that captures the agent’s knowledge about the environment and about the other agents, including its knowledge about their states of knowledge, which can include what they know about the other agents, and so on. This approach has been called the decision-theoretic approach to game theory. It avoids some of the drawbacks of game-theoretic equilibria that may be nonunique and do not capturo off-equilibrium behaviors, but it does so at the cost of having to represent, process and continually update the nested state of agent’s knowledge.