What is it about?
We show how exploiting the information about the variables interaction of an objective function it is possible to apply efficient methods to solve the optimization problems (if the problems are easy). Traditional deceptive functions can be easily solved this way. On the other hand, the hard problems we can find in an industrial context has some structure that is missing in the random instances used in the academic papers. This suggests that we could exploit this structure to improve the optimization algorithms for industrial instances.
Featured Image
Why is it important?
The describes what is easy or difficult to do in pseudo-Boolean (binary) optimization. The reader can find some surprises inside, and find easy problems that are considered "hard" for many methods.
Perspectives
I think this is work all the researchers interested in pseudo-Boolean (binary) optimization should read, since it revises some optimization methods and ideas that should be present in the mind of researchers when designing new optimization algorithms.
Dr. Francisco Chicano
University of Málaga
Read the Original
This page is a summary of: Gray Box Optimization for Mk Landscapes (NK Landscapes and MAX-kSAT), Evolutionary Computation, September 2016, The MIT Press,
DOI: 10.1162/evco_a_00184.
You can read the full text:
Contributors
The following have contributed to this page
