Distributed resource coordination strategies for mobile ad hoc networks

2006 2006

Other formats: Order a copy

Abstract (summary)

This work was motivated by the distributed information processing needs in tactical environments such as disaster relief and military operations. Tactical network environments are characterized by mobile ad hoc networks under resource and policy constraints, tasked for critical missions.

Mission success directly depends on the effective use of computation resources in the field for distributed data processing and information propagation. However, the dynamic nature of the network and the lack of centralized coordination components make it very difficult to globally allocate and maintain resources for distributed tasks in a manner that is itself distributed, efficient and adaptive to the volatile nature of the environment.

Current approaches to the problem can be broadly classified in three main categories, centralized decision making (applicable only to small scale networks), local greedy decision making and arbitrage models also known as agent-based negotiation.

In this work, we introduce a new solution to the problem which utilizes online learning strategies at the local node level, to quickly evolve the global resource allocation solution that asymptotically converges to a global optimum.

The resource allocation problem in mobile ad hoc networks is first formulated as a k-arm bandit problem at the local level. As data flows through the network, each node locally learns the best policies to the used under different data flows, different constraints and local network topology.

Two learning strategies (ϵ -greedy and SoftMax) are adapted to the problem domain and used for tests and comparisons. A proof-of-concept implementation of the proposed resource allocation algorithm is introduced, discussed and tested in simulated networks.

The preliminary experimental results and the theoretical guarantees provided for the algorithm indicate that the approach is applicable to the resource allocation problem in mobile ad hoc networks for tactical environments.

Indexing (details)

Computer science
0984: Computer science
Identifier / keyword
Applied sciences; Mobile ad hoc networks; Reinforcement learning; Resource coordination
Distributed resource coordination strategies for mobile ad hoc networks
Carvalho, Marco M.
Number of pages
Publication year
Degree date
School code
DAI-B 68/07, Dissertation Abstracts International
Place of publication
Ann Arbor
Country of publication
United States
Buckles, Bill
Tulane University School of Science and Engineering
University location
United States -- Louisiana
Source type
Dissertations & Theses
Document type
Dissertation/thesis number
ProQuest document ID
Database copyright ProQuest LLC; ProQuest does not claim copyright in the individual underlying works.
Document URL
Access the complete full text

You can get the full text of this document if it is part of your institution's ProQuest subscription.

Try one of the following:

  • Connect to ProQuest through your library network and search for the document from there.
  • Request the document from your library.
  • Go to the ProQuest login page and enter a ProQuest or My Research username / password.