Tutorial on Modern Linkage Learning Techniques in Combinatorial Optimization




The 2022 year updates in this tutorial are marked in red.





Aim and scope of the tutorial






Linkage learning is employed by many state-of-the-art evolutionary methods dedicated to solving problems in various domains: binary, discrete non-binary, permutation-based, continuous, and others. It has been successfully applied to solving single- and multi-objective problems. The information about underlying problem structure, discovered by linkage learning, is the key part of many state-of-the-art evolutionary methods. However, linkage learning techniques are often considered hard to understand or difficult to use. Therefore, the aims of this tutorial are as follows:

-          Present the current research state concerning linkage learning (more attention paid to the recent propositions concerning Empirical Linkage Learning, including the Dark Gray-box optimization)

-          Present in detail, using easy-to-understand examples, the work of modern linkage learning techniques

-          Present in detail, using easy-to-understand examples, how linkage may be represented and utilized

-          Present the details of state-of-the-art evolutionary methods that employ linkage learning

-          Present the differences between linkage learning application in single- and multi-objective optimization problems

-          Present how linkage quality affects the evolutionary search

-          Present the most important challenges that are currently faced by linkage learning development

-          Present the most promising future work directions for research that concerns linkage learning


Linkage learning techniques apply to any optimization domain. However, linkage learning techniques dedicated to continuous search spaces are usually significantly different than those proposed for combinatorial problems. Therefore, this tutorial will focus on linkage learning techniques dedicated to discrete (including  binary) and permutation-based search spaces. Nevertheless, for the presented techniques, we will point to their successful applications in continuous search spaces. The main utility features of the tutorial will be as follows.

-          All parts of the tutorial will be based on simple examples that will picture the presented issues

-          The main aim of the tutorial is to make the issues related to modern linkage learning easy-to-understand

For the tutorial we will prepare a source code pack (accessible via GitHub), so the participants will be able to test the discussed techniques and methods on their own




Tutorial length

2 hours







Tutorial level















See tutorial overview (agenda)






See tutorial organizers