org.opentripplanner.analyst.core

Class WeightingFunction.Logistic

java.lang.Object

org.opentripplanner.analyst.core.WeightingFunction

org.opentripplanner.analyst.core.WeightingFunction.Logistic

Enclosing class:

WeightingFunction

public static final class WeightingFunction.Logistic
extends WeightingFunction

The logistic function, a commonly used sigmoid (s-shaped function). Basically a step function with a smooth rolloff. Positive values of k cause the function to increase from zero to one, negative values will be inverted.
k=-1 / 60 gives a rolloff over an interval of ~10 minutes
k=-0.5 / 60 gives a rolloff over an interval of ~20 minutes
k=-2 / 60 contracts the rolloff interval to ~5 minutes

Nested Class Summary

Nested classes/interfaces inherited from class org.opentripplanner.analyst.core.WeightingFunction

WeightingFunction.Logistic, WeightingFunction.SharpCutoff

Constructor Summary

Constructors

Constructor and Description
Logistic(double steepness)

Method Summary

Modifier and Type	Method and Description
int[]	apply(int[] countsPerSecond) Apply the logistic weighting function.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Logistic

public Logistic(double steepness)

Method Detail

apply

public int[] apply(int[] countsPerSecond)

Apply the logistic weighting function. Uses some slightly clever optimizations. We have precomputed the weights for the sigmoid centered on 0 and stored them in the weights array, offset by -rolloffMin. We loop over the values to output and first find the largest second where the weight is effectively 1. We store a cumulative sum up to that point. We then run through the next

Specified by:

apply in class WeightingFunction