iEvolve: Incentive mechanisms for evolutionary topology formation in autonomous networks with strategic agents.

摘要

In an autonomous network with strategic agents, cooperations among nodes cannot be assumed, since each node is capable of making independent decisions based on their personal preferences. Ideally, are incentive-compatible mechanism should be provided to motivate cooperations among autonomous nodes so that a mutually beneficial network results. In this thesis, we apply mechanism design to solve two problem scenarios in autonomous networks, namely the evolutionary topologies problem and the overlay multicast problem. In each scenario, we designed a distributed incentive mechanism that motivates each node to evolve into a more desirable network topology. We name our incentive-compatible distributed algorithms iEvolve. The correctness and performance of each distributed algorithm are verified by simulation and implementation.