1.1 Research Motivation and Purpose
Multi-hop cellular networking has been an active research area in recent years. In conventional cellular networks, mobile stations communicate directly with their assigned base station; in wireless multi-hop networks, mobile stations are located randomly and use peer-to-peer communications to relay their messages. Multi-hop cellular networks that integrate the characteristics of both cellular and mobile ad hoc networks to leverage the advantages of each other have received increasing attention. Figure 1.1 indicates the scenario of general multi-hop cellular networks, the service area of the cellular networks can be extended by adopting hop-by-hop connections at the boundaries of the cell.
Figure 1.1 Scenario of general multi-hop cellular networks
Much research has evaluated and summarized the benefits of such a hybrid architecture [1][4][7][8]:
z The energy consumption of the mobile device can be conserved.
z The interference with other nodes can be reduced.
z The number of fixed antennas can be reduced.
z The capacity of the cell can be increased.
z The coverage of the network can be enhanced.
z The robustness and scalability of the system can be increased.
In multi-hop cellular networks, data packets must be relayed hop by hop from a given mobile node to a base station and vice-versa [8]. Cooperation among nodes is an important prerequisite for the success of the relaying ad-hoc networks. In cooperating groups, such as emergency and military situations, all nodes belong to a single authority and therefore have a good reason to support each other. However, in the groups of anonymous participants, such as emerging civilian applications, the nodes do not belong to a single authority and cooperative behaviors can not be directly assumed [15]. Moreover, forwarding data for others incurs the consumption of battery energy and the delay of its own data, the assumption of spontaneous willingness to relay data is unrealistic for autonomous mobile nodes [16]. Consequently, providing incentives for the mobile nodes to cooperate as relaying entities in the groups of mutually unknown participants is reasonable.
Much research [7-8, 15-22] has described how to stimulate intermediate nodes to forward data packets in multi-hop networks. The approaches can be classified into detection-based and motivation-based. The detection-based approach finds out misbehaving nodes and mitigates their impact in the networks. The motivation-based approach provides incentives to foster positive cooperation in ad hoc networks. Most works on the motivation-based approaches focus on its protocol and security aspects or just employ fixed-rate pricing on number of packets or volume of traffic forwarded. The major advantage
of the fixed-rate pricing is that billing and accounting processes are simple. However, the price of the feedback incentives is independent of the actual state of the networks. Such system cannot react effectively to the dynamic and unpredictable variations of the wireless networks.
Cost savings and service availability are two major concerns of a network provider adopting multi-hop cellular networking technology. In this research, we propose a dynamic incentive pricing scheme to maximize the revenue of the network provider while maintaining service availability in the networks. Monetary incentives not only affect the motivation of the intermediate nodes supporting relaying services but represent the costs of providing connection services in multi-hop cellular networks. If the price of the incentives is too low, the number of successful connections will be small and the network provider can not get adequate profit from the relaying networks. However, if the price of the incentives is too high, the network provider can not cover the costs from the fee charged from end users.
Consequently, dynamically adjusting the price of the incentives based on the network conditions is more appropriate than fixed-rate pricing for the network provider to make maximum revenue.
Since providing a uniform price of the incentives to all mobile nodes depending on the network situations neglects the distinct importance of each mobile node in the network topology, we also investigate how to give different amount of incentives for each mobile node that has different effects on supporting hop-by-hop connections. The base station should give more incentives to the nodes of high importance so that it can make more mobile nodes connect to the base station successfully. In order to react effectively to the individual impact of each mobile node on service availability, we present the concept of location-based incentive pricing for relaying services in multi-hop cellular networks. When a pre-constructed routing topology exists in the networks, the price of the feedback incentives for each
intermediate node is adjusted according to the number of nodes that reside in its sub-tree. The willingness of an intermediate node relaying packets has a significant impact on the success of the multi-hop connections from all nodes in its sub-tree to the base station. The proposed pricing scheme shifts incentives from the nodes of low importance to the nodes of high importance in the routing topology so that it increases service availability without additional costs.
When a pre-constructed routing topology is not available in the networks, a new metric called Quality-of-Relay (QoR) is defined to evaluate the importance of each mobile node affecting other nodes that require hop-by-hop connections to reach the central base station.
Shifting incentives from the nodes of low importance to the nodes of high importance in the networks also enhances service availability with only a slight increase in relaying costs. In addition to adopting the QoR value for incentive pricing, we present a routing scheme based on the QoR value of each mobile node in the networks to select an optimal relaying path for connecting to the central base station. Although shortest path is the most simple and common metric used in the routing protocol, it may route almost packets over a few paths and result in network congestion and resource unavailable in hot spot. The routing scheme that selects a relaying path with minimum sum of the QoR values of all intermediate nodes in the path can retain more valuable resource for later relaying requests, thereby accepting more relaying connections under a certain constraint on maximum relaying capacity of each mobile node.
1.2 Outline of This Thesis
This thesis is organized into six chapters. Chapter 1 outlines the motivation and the research purpose. In Chapter 2, we review the existing multi-hop cellular network models and incentive approaches. Then, we introduce the general supply function for providing relaying services. In Chapter 3, a dynamic incentive pricing scheme is presented to maximize the
revenue of the network provider by adjusting the price of the feedback incentives based on the actual network conditions. Chapter 4 proposes a location-based incentive pricing scheme for relaying services with a tree-based routing topology. Chapter 5 describes Quality-of-Relay based incentive pricing and routing for relaying services without a pre-constructed routing topology. The pricing schemes presented in Chapter 4 and 5 encourage collaboration depending on the degree of each mobile node contributing to successful hop-by-hop connections. Finally, we conclude our research and suggest possible further research directions in Chapter 6.