2011 Fiscal Year Final Research Report
Autonomous decentralized control based on local interaction for mobile ad hoc networks
| Project/Area Number |
21700086
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
Computer system/Network
|
|---|
| Research Institution | Hiroshima City University |
|---|
Principal Investigator |
TAKANO Chisa 広島市立大学, 情報科学研究科, 准教授 (60509058)
|
|---|
| Project Period (FY) |
2009 – 2011
|
| Keywords | モバイルネットワーク技術 / 自律分散トラヒック制御 |
|---|
| Research Abstract |
The aim of this work is that we apply "the autonomous decentralized control based on the local interaction" considered so far to an ad hoc network, and realize network resource management by which the state of the whole network is controlled indirectly through the autonomous operation of each node based on local information.
For the research plans(Subject 1-Subject 3), our achieved contents are as follows:
(Subject 1) Problems of destruction of cluster structure for the dynamic network topology
This method does not guarantee asymptotic stability with regard to the size of the amplitude of the distribution that represents the cluster structure. Therefore, it is difficult to keep the structure in a dynamic environment. In this study, we considered and evaluated two approaches to solve this problem. In the first approach, we introduced an amplification operation to guarantee the asymptotic stability. However, we verified that amplitude changes exponentially even if the amplification factor d
… More
eviates only slightly from the ideal value. Thus, this approach is rather impractical in real networks. On the other hand, our second approach is a method that uses distribution vectors to preserve distribution history. This method can control the number of clusters and cluster size by changing the components of the distribution vector. In addition, it is possible to guarantee the stability of the amplitude of the distribution. As future work, we will examine the temporal variation of the amplitude of the distribution in case that the initial state of the distribution changes over time.
(Subject 2) Problems of simplification of a network model
We assumed the limited network model like a two-dimensional lattice-like network model in which nodes did not move. In this research, we evaluated the effects of the proposed methods and confirmed the validity in consideration of movement of nodes by using the complicated network model like unit graph.
(Subject 3) Evaluation of network performance to the proposed clustering method
We made the evaluation index the battery residual quantity of each node concretely, and evaluated the FND time(First Node Die time : time until the first node becomes impossible to use in ad hoc network) and the temporal evolution of the probability of survival nodes at the time of a cluster configuration. As a result, we showed that the proposed method was superior to the existing clustering method. Less
|
