A Novel Game Based Incentive Strategy for Opportunistic Networks

Opportunistic networks are a emerging networks characterized by frequent network partitions, high bit error ratio and random topology instability, where the message propagation depends on the cooperation of nodes to fulfill a "store-carry-forward" fashion. Due to the constrained energy, memory and processing capacity, some individual nodes may behave selfishly, or even maliciously, which will introduce damage into the existing routing schemes based on cooperation and degrade the performance (lower delivery ratio, longer latency etc.,) of opportunistic networks greatly. In order to address the above issues, the current price-based incentive strategy, Credit relies on a fixed management-center which is rare in the realistic opportunistic networks with little infrastructure to manage the transaction that the source of messages pays virtual credits to nodes that relay messages for it. So this paper proposes a novel Game based Incentive Strategy (GIS) which utilizes three-time bargaining model based on two-person transaction and allows the sending nodes to pay the relay nodes directly according to the optimal price drawn by game without any third party. GIS stimulates the cooperation of selfish nodes to forward messages effectively, while holds back the deceptive price stemmed from the malicious intermediary nodes to facilitate deals. From the extensive simulations results, GIS can optimize the average latency and the delivery ratio to the greatest extent. Additionally, effectiveness and fairness can be guaranteed.

I. Introduction

Generally speaking, some classical routing algorithms [1] [2] [3](e.g., Epidemic, Prophet and Spray-Wait) work with a premise that each node unit were the volunteer for assisting others to forward messages in opportunistic networks [4] [5]. Unfortunately, the assumption doesn't hold in the actual wireless network environment. Specifically, few nodes are selfless to act in connect with others unconditionally when the devices with the confined processing power, buffer, and battery capacity are controlled by people or other rational entities. Thus the active cooperation between the nodes can not be taken for granted. The vast majority of studies demonstrate that they will also lead to a significant decrease [6] in network function, even though there exist only a small number of non–cooperation entities. So many scholars are eyeing this object so as to beef up the lagging networks experience for users. Currently, the available schemes can be classified into three main categories: reputation systems [7], tit-for-tat (TFT) [8] systems and price-based systems. The above are mainly concentrated on the selfish individual rather than the malicious one on cutting edge topic. Such as Credit strategy [9], disseminating messages can be modeled as a transaction based on the virtual currency and a fixed management–center which managers the transaction is indispensable. So it just single–mindedly pursues the cooperation of selfish nodes by money–paid mode instead of preventing the deceptive price from the malicious individual ascribed to the third party's management that sending nodes pay all nodes with the same credits.