The files were used in order to create the results for the paper Bivariate Estimation-of-Distribution Algorithms Can Find an Exponential Number of Optima.
They act as a means to clarify implementation details of the algorithm, the objective function, as well as the test setup.
However, the files are not necessarily intended to be used out of the box, although all files with a main can be ran (that is, all files but fitness_functions.d, mimic.d, and mimic_looking_for_more_optima.d), resulting in data that was used in the paper.
The code is written in D, for which several compilers exist. Any compiler that works for your system is well suited.
Important: Running experiments that save all of the different optima found during the optimization process (number_of_different_optima.d) results in massive files (about 90 GB!), depending on the dimension size of the problem.