Stochastic universal sampling

<h2 id="pseudo-code">Pseudo Code</h2>
Described as an algorithm, pseudocode for SUS looks like:

SUS(Population, N)
 F := total fitness of Population
 N := number of offspring to keep
 P := distance between the pointers (F/N)
 Start := random number between 0 and P
 Pointers := [Start + i*P | i in [0..(N-1)]]
 return RWS(Population,Pointers)

RWS(Population, Points)
 Keep = []
 for P in Points
 I := 0
 while fitness sum of Population[0..I] < P
 I++
 add Population[I] to Keep
 return Keep

Where Population[0..I] is the set of individuals with array-index 0 to (and including) I.
Here RWS() describes the bulk of fitness proportionate selection (also known as "<a href="/facts/Roulette_wheel_selection/RSeQxJXS">roulette wheel selection</a>") – in true fitness proportional selection the parameter Points is always a (sorted) list of random numbers from 0 to F. The algorithm above is intended to be illustrative rather than canonical.

<h2 id="see-also">See also</h2>
<ul><li><a href="/facts/Fitness_proportionate_selection/RSeQxJXS">Fitness proportionate selection</a></li>
<li><a href="/facts/Reward-based_selection/5MqJgyxD">Reward-based selection</a></li></ul>

<h2 id="references">References</h2>

<ol>
<li id="fn:1">Baker, James E. (1987). "Reducing Bias and Inefficiency in the Selection Algorithm". Proceedings of the Second International Conference on Genetic Algorithms and Their Application. Hillsdale, New Jersey: L. Erlbaum Associates: 14–21. <a href="#fnref:1" class="footnote-back-ref">↩</a></li>
</ol>

Stochastic universal sampling open-in-new

Stochastic universal sampling