This paper contributes to the analysis of the question how altruistic behaviour can be in an agent’s own interest. The question is addressed by involving a temporal dimension in which altruist behaviour at one point in time can be in the agent’s interest at a future point in time, depending on the environment. The claim is that to be able to make reasonable decisions, an agent needs a cognitive system for intertemporal decision making, in relation to a model of the environment. To address this, a society of agents has been modelled, simulated and analyzed. Some of the agents have a cognitive system including a model of the environment based on a dynamic model for trust in other agents. This environment model is combined with a model for intertemporal decision making. Agents with this cognitive system show more altruistic behaviour, they get a larger social network, and become in the end more healthy than agents without such a cognitive system.
Tibor Bosse, Martijn C. Schut, Jan Treur, David We