Utility Fair Flow Control In Wireless Sensor Networks

The low-lost, strong dynamic, self-organization and high fault tolerancecharacteristics of wireless sensor networks make it have broad application prospectsand development space. Recently, the research of wireless sensor networks have madea rapid development, however, as the inherent limitations such as computation,storage and communication abilities, how to design effective protocols and algorithmsto complete various tasks efficiently with limited resources has become a majorchallenge in wireless sensor networks.Based on the full analysis of overseas and domestic research status on resourceallocation and flow control in wireless sensor networks, in view of the limitedresources, diverse businesses in wireless sensor networks and traditional subgradientalgorithms always have defects such as slow convergence rate and largecommunication overhead, we mainly do the following research contents andinnovation works in this paper:(1) We succeed propose flow control problems with joint constraints of sensornodes’ receiver capacity and energy in wireless sensor networks and rechargeablewireless sensor networks. Compared with traditional flow control problemformulations, the innovation points and advantages of flow control problemformulations proposed in this paper mainly reflects in:①We adopt the receivercapacity model, which could overcome the influence of interference links to linkcapacity and is more close to the actual network scene.②Compared with traditionalrate fair flow control mechanism, the utility fair flow control strategy introduced inthis paper could support heterogeneous communication and could suitable for a mixof elastic and inelastic traffic.③We propose distributed event-triggered algorithmsfor the utility fair flow control problems, which could reduce the communication costsand facilitate the implementation in large-scale network scene.(2) As present research efforts mainly exist two defects as:①The flow controlmechanism adopted only suitable for elastic traffic (modeled by strictly concaveutility function),②The traditional dual decomposition and subgradient algorithmsadopted often have defects such as slow convergence rate, difficult adjustment ofstepsize and large communication overhead, we first investigated the flow control problem for mixed traffic in wireless sensor networks. Considering joint constraints ofreceiver capacity and network lifetime, we design a distributed event-triggeredalgorithm which could effectively solve the utility fair flow control problem inwireless sensor networks. The effective solution could efficiently supply resourceallocation for both elastic and inelastic traffic, reduce the network communicationcosts and facilitate the implementation in large-scale network scene.(3) Extending to rechargeable wireless sensor networks with limited batterycapacity, we investigate the utility fair flow control problem with joint constraints ofreceiver capacity and energy allocation in rechargeable wireless sensor networks.Based on optimal energy management strategy, in order to remove the couplingbetween flow rates and energy allocation, we adopt the adaptive energy allocationscheme to determine the effective energy allocation for sensor nodes, leading to asequence of independent flow control subproblems that admit distributed solutionsamong different time slots. With adaptive energy allocation scheme, the energyallocated for sensor nodes are dynamic in different time slots, resulting in therelationship of sensor nodes’ receiver capacity and energy allocation for the utility fairflow control varies in different time slots, thus we provide a heuristic but effectivepenalty parameter adjusting method to solve each flow control subproblem efficientlywith distributed event-triggered algorithm.