/topology_preserving_maps

Primary LanguageJupyter Notebook

Topology Preserving Maps

Input: 1-hop neighbourhood information for every node, Geographic Coordinates of Perimeter Nodes.
Output: Topological Coordinates of all the nodes, which closely resemble the geographic map.
As an intermediate Step, this algorithm also produces the virtual coorinates of all the nodes in the network.

How to run:

  1. Execute the python3 calculate_virtual_coordinates.get_VC(num_of_nodes) function in the file backup/TPM_from_VCS/util/calculate_virtual_coordinates.py. This function makes use of the other functions in defined in that file.

  2. Then execute the function generate_tpm_from_vcs(VC_Matrix) in the file backup/TPM_from_VCS/util/tpm_from_vcs.py. This produces the Topological Cordinates as proposed in the below paper.

Info about other folders:

  1. backup/TPM_from_VCS/data/ contains the programs to generate synthetic datasets.
  2. backup/TPM_from_VCS/models/ contains different RNN models for generating Topological Coordinates.
  3. backup/TPM_from_VCS/plot_functions/ contains generic matplotlib funtions to visualise the output of the algorithm.

*.meta and *.index are tensorflow checkpoint files

The below files are works in progress:

  1. backup/pg_1.ipynb : A Policy Gradient approach
  2. backup/ae-basic-Copy1.ipynb: Auto Encoder Approach

References

  1. Dhanapala, Dulanjalie C., and Anura P. Jayasumana. "Topology preserving maps: extracting layout maps of wireless sensor networks from virtual coordinates." IEEE/ACM Transactions on Networking (TON) 22.3 (2014): 784-797.