Select topics in the analysis of evolutionary algorithms

Abstract

Now that we have a collection of analytical tools tailored for the analysis of evolutionary algorithms, we are ready to consider concrete evolutionary algorithms and derive results, gain insights, prove facts about their behavior and performance. Of course, what one finds interesting and worth investigating in evolutionary computation is a personal question that is a matter of taste and other circumstances. It is almost inevitable that some questions that a reader finds most interesting will not be considered here. Therefore, the most important purpose this chapter serves is to be an example of how evolutionary algorithms can be analyzed using the methods described in the previous chapter. We do this considering four different topics that cover four different aspects of evolutionary computation. This includes considering effects of specific variation crossovers, dipping into the topic of design of evolutionary algorithms, considering a specific variant of evolutionary computation that we have not covered before, and, finally, considering an example for an application of evolutionary algorithms in combinatorial optimization.