Altruistic agents in dynamic games

The collective effort of the agents that operate distributed, dynamic networks can be viewed as a dynamic game. Having limited influence over the decisions, the agents react to one another's decisions by resolving their designated problems. Typically, these iterative processes arrive at attractors that can be far from the Pareto optimal decisions those yielded by an ideal, centralized agent. Herein, the focus is on the development of augmentations for the problems of altruistic agents, which abandon competition to draw the iterative processes towards Pareto decisions. This paper elaborates on augmentations for unconstrained, but general problems and it proposes an algorithm for inferring optimal values for the parameters of the augmentations.