Bin Smoothing

From Robowiki
Revision as of 04:46, 9 June 2009 by Nat (talk | contribs) (cleanup)
Jump to navigation Jump to search
This article is a stub. You can help RoboWiki by expanding it.

Bin Smoothing is the process of spreading or normalizing the values stored in a Bin for Wave Surfing or Segmentation. It's not needed for Wave Surfing or Segmentation, but can, in theory, improve performance if done right.

How Bin Smoothing works

Bin Smoothing 'flattens' small spikes within it's Bins.

Bin Smoothing Formulas

If you have your own Bin Smoothing Formula, feel free to post it here.

Pugilist's

    double smoothedVisits(int index) {
	double smoothed = 0;
	int i = 0;
	do {
	    smoothed += (double)visits[i] / Math.sqrt((double)(Math.abs(index - i) + 1.0));
	    i++;
	} while (i < Pugilist.FACTORS);
	return smoothed / Math.pow(distanceToTarget() / bulletVelocity, 1.3);
    }

PPP

double smoothedVisits(int index) {
    double smoothed = 0;
    if (index > 0) {
        smoothed += visits[index - 1] / 2;
    }
    if (index < FACTORS - 1) {
        smoothed += visits[index + 1] / 2;
    }
    smoothed += visits[index];
    return smoothed;
}

Other ways

Math.pow(0.5, Math.abs(index - count))

This does a 50% decrease.

Math.pow(0.1, Math.abs(index - count))

This does a 90% decrease

1.0/(double)(1 + (index - count)*(index - count))