Mobile Adhoc Offloading

This problem is a series of biddings and auctions. Each round of bidding and auction are different from previous ones because of the change of network topology, variance of budget set by the sender, and possible evolution of strategies of other nodes. The huge strategy space of relay nodes makes the formulation to a game very difficult. We present a brief qualitative analysis in this paper, and propose a bidding strategy based on learning algorithms. I. PRELIMINARIES AND NOTATIONS In this paper, Section I to III describe proposed scheme which competed in MANIAC competition, section IV briefly reports the performance of this scheme, lessons learnt from the competition is concluded in section V. Both AP and Handhelds auction off the traffic load forwarding service with the same elements: budget, fine, time out, source and destination. As the forwarding task has no information of the auctioneer in it, thus downstream entity 1 adapts strategies without care the auctioneer being AP or handheld. For convenience of analysis, we generalize a recursive auction process as shown in following figure,