Research On Data Dissemination And Collection Methods In Mobile Opportunistic Networks

Mobile opportunistic networks(MONs)can be applied to the scenario of frequent interruption of communication links and lack of communication infrastructure,and meet the requirements of ad hoc network in the actual scenario.MONs have a significant impact on the popularization of Internet of things technology in the future.Data dissemination and collection are the basis of many application in MONs and are of great research significance.However,the characteristics of intermittent connectivity between nodes,limited node resources and long delay in data transmission have brought great challenges to the design of efficient data dissemination and collection schemes.Although there are a lot of researches on data transmission in MONs,rnost of them only focus on the"one-to-one" transmission form,that is,the transmitted message has only one.source node and destination node.There are still many problems to be solved for "one-to-many" data dissemination and "many-to-one" data collection,such as how to reduce network overhead in data dissemination scheme,how to realize efficient data dissemination without clearly specifying message receiver,and how to collect data efficiently under delay constraints.Thus,in this thesis,we focuses on the problem of data dissemination and collection in MONs,and studies from three aspects:data dissemination scheme for node IDs,data dissemination scheme for node attributes and data collection strategy satisfying delay constraints.We summarize our main contributions as follows:(1)Data dissemination scheme based on network layering in MONs.Given a set of destination node IDs,multicast is an important means of data dissemination.Inspired by the use of backbone to reduce network overhead and improve routing performance in traditional network,this thesis proposes a load-balanced backbone-based multicast routing algorithm in MONs.Firstly,we transform the construction of virtual multicast backbone in MONs into an optimization problem with the optimization objectives of maximizing the ability of a backbone to transmit multicast messages,minimizing the inter-contact time between backbone nodes and their neighbor nodes,and the load balancing of nodes in the backbone.We propose a heuristic algorithm based on multi-objective optimization to solve the problem.Secondly,using the node’s residual buffer and the local multi-hop neighbor node information maintained by the node,we design a local backbone maintenance algorithm.The algorithm selects nodes in the MON to replace the backbone nodes with insufficient buffer,which increases the robustness of backbone-based multicast routing algorithm when the buffer of backbone nodes is insufficient.(2)Profile-cast in MONs.Profile-cast allows messages to be sent for node attributes rather than node IDs.Most of the existing works focus on the behavior attributes or interest attributes of nodes.In these works,it is usually assumed that nodes with similar attribute have a high encounter probability,and the transmission of messages is limited to nodes with similar attribute.These works only consider a subset of scenarios in profile-cast.When designing the scheme,this thesis decouples the node’s attribute values and its encounter behavior,and proposes a profile-cast scheme that can be applied to more scenarios.Firstly,in the scenario with a fixed source node,we propose an algorithm based on message forwarding tree,which aims to minimize the delay of messages and reduce the overhead of the profile-cast scheme.When constructing the message forwarding tree,we use a heuristic algorithm based on tabu search to select the terminal node of the forwarding tree.Secondly,in the scenario with unfixed source node,we propose a distributed profi le-cast scheme,which depicts the frequently encountered nodes according to the encounter frequency between the node and its neighbor nodes and the neighbor nodes’ attribute information,so as to limit the transmission range of messages.To reduce the overhead of profile-cast scheme,we use network coding to reduce the communication overhead between nodes and their neighbor nodes.(3)Data collection under dual delay constraints in MONs.This thesis transforms the problem of data collection by nodes with fixed moving trajectory in MONs into the problem of sweep coverage of mobile nodes.Different from the sweep coverage problem in traditional WSNs,in MONs,in addition to the data collection period constraint,data transmission delay needs to be considered.Under the dual delay constraints of data collection points,this thesis designs a node movement path generation algorithm,which aims to cover all collection points in the collection region with the least mobile nodes.When the collection period constraints of different data collection points are different,we design a method to divide the set of collection points to expand the application scenario of our proposed algorithm.In order to expand the range of collection points that can be covered by mobile nodes,we design an algorithm for nodes with different trajectories to cooperate with each other to transfer data.In summary,this thesis studies the problem of data dissemination and collection methods in MONs,and proposes a series of methods.We also verify the proposed models and methods through theoretical analysis and extensive experiments.The work of this thesis provides theoretical and technical support for the wide application of mobile opportunistic network.