hi-opt: Human Intranet Optimization Platform

Requirements

• Castalia Network Simulator; available at: https://github.com/boulis/Castalia
  • Castalia is based on OMNeT++ simulator, so you need to install it first as discussed in Castalia's installation manual.
  • The current version of Castalia only works with OMNeT's older version 4.x (been tested with 4.6).
• Python
• Pulp LP modeler Python package; available at: https://pypi.python.org/pypi/PuLP
• IBM ILOG CPLEX Optimization Studio with its Python API

Installation

• Install Castalia, Python, Pulp, and CPLEX.
• The following folders should be added to the default Castalia installation. You can either replace the whole "Castalia-3.2" folder with the one provided in this repo, or copy only these folders if you already had Castallia installed with your own settings:
  • Castalia-3.2/src/node/application/hiMesh
  • Castalia-3.2/src/node/application/hiStar
  • Castalia-3.2/src/node/communication/mac/hiTdmaMac
  • Castalia-3.2/out/gcc-release/src/node/application/hiMesh
  • Castalia-3.2/out/gcc-release/src/node/application/hiStar
  • Castalia-3.2/out/gcc-release/src/node/communication/mac/hiTdmaMac
  • Castalia-3.2/Simulations/HIBAN
  • Castalia-3.2/Simulations/Parameters/Radio/CC2650.txt

Package Structure

• all_cases_run.py: Running this script generates the results for all feasible cases, and saves them in a file (all_results.txt).
• anneal.py: Simulated annealing algorithm for comparison (package available at https://github.com/perrygeo/simanneal)
• configs.py: Main configurations to be set by user.
• exhaustiveSearch.py: Reads all cases from all_results.txt, sorts them, and prints the optimized results for each PDRmin.
• funcs.py: Miscellaneous internal functions for running simulations and calculating statistics.
• main.py: Main optimization algorithm.
• patloss.py: The average path loss values can be changed here. The script generates the requierd pathLossMap.txt for Castalia for each simulation run.

Running Optimization

• Open configs.py and specify the following parameters:
  • castalia_path: Your Castalia's installation path.
  • PDRmin: Minimum PDR constraint.
  • inputFileMode: There are two modes for running the optimizer:
    • inputFileMode = False: The full algorithm will be run with the required simulations. This can take a long time depending on the input constraints.
    • inputFileMode = True: This mode is useful when you want to try different constraints on the same problem. In this case, you have to first run all_cases_run.py script which simulates all feasible cases and saves them in all_results.txt. This will take a very long time. However, the optimization algorithm will look up through this file from now on instead of running simulations each time. Note: all_results.txt is provided to you for the sample problem discussed in the paper, so you don't need to run all_cases_run.py for that problem.
• Run main.py. The optimal solution would be printed on the screen.

Problems?

If you face any problems or discover any bugs, please let us know: MyLastName AT berkeley DOT edu

For more info, you can read and cite our paper:

A. Moin, P. Nuzzo, A. L. Sangiovanni-Vincentelli, and J. M. Rabaey, "Optimized Design of a Human Intranet Network," 2017 54th ACM/EDAC/IEEE Design Automation Conference (DAC), Austin, TX, 2017.