In this work, multimodal test problems are generated 
for comparison of different optimization algorithms 
via a property based scheme, including the recently 
introduced features of global-local contrast and 
basin size contrast. The goal is to compare how 
two niching-based evolutionary algorithms cope 
with these problems. In order to detect similarities 
and differences, Response Surface Models over the 
problem features are employed. Interestingly, the 
problem features lead to a qualitatively similar 
behavior of the algorithms, however, the obtained 
performance values are very different and suggest 
that one algorithm is more focused on detecting 
all basins, and the other rather on pursuing the 
global optimum.