Just Use Distributions: Eliminating The Noise of Sampling-Based Fitness
Estimation
- Robert McKay ,
- zhenyue Qin ,
- Tom Gedeon
Abstract
This paper uses distributional analysis to disentangle the effects of
fitness landscape and stochasticity in Wagner's Gene Regulatory Network
Based evolutionary scenario. Further advantages of the analysis include
the ability to accurately identify global optima and potential
extensions to approximate the distributional fitness more accurately
than stochastic sampling at lower computational cost.
This is an inital draft, and will be updated as further content is
written.