Week ending 30th April

The whole week was spent in coding and testing the new algorithm I developed from the previous week's research. The SpringEq algorithm I call.

I am testing the new algorithm

• In landmark placement scenario
• Test for the convergence speed (no of iteration), computational speed, and accuracy.
• Tested with synthetic network and all-to-all Planetlab ping data
• When using Planetlab data smoothing is required. Did some review and testing of existing approaches (absolute average, moving average, exponential weighted average, and double exponential weighted average) and selected the exponential smoothing for my purpose.

The testing is not completely finished.

Also by the end of the week, I realized the need for comparing the simplex and spring algorithms for the node placement application. I am doing that too now.