Social welfare of selfish agents: motivating efficiency for divisible resources

In today's landscape of distributed and autonomous computing, there is a challenge to construct mechanisms which can induce selfish agents to act in a way that satisfies a global goal. In the domain for the allocation of computational and network resources, proportionally fair schemes are commonly advocated. In this paper, we investigate the efficiency of the resulting equilibria in such systems. We then develop a method of generating an entire class of divisible auctions with minimal signaling and computation costs which maximize social welfare even though agents act solely to optimize their own utility.