The NK-IFP Benchmarks

As described in the article, the benchmark are a combination of two NKL Models, F^A(x) and F^B(x), with different K and different neighbourhood definitions by a simple multiplication:

Currently two benchmarks have been designed.

NK-IFP-1

F^A(x): a K = 4 semi-adjacent circular neighborhood is designed as follows: {x_i−2, x_i−1, x_i+1, x_i+2}, omitting the central position x_i.
F^B(x): a K = 3 neighborhood of uniform random distribution.

NK-IFP-2

F^A(x): a K = 4 semi-adjacent circular neighborhood as NK-IFP-1.
F^B(x): a K = 5 neighborhood of uniform random + 20 positions wide triangular distribution.