TITLE: Analysing Mutation in Randomised Search Heuristics
SPEAKER: Thomas Jansen (Aberystwyth University)

ABSTRACT:
Randomised search heuristics are a huge class of general-purpose problem solvers that encompass many nature-inspired methods like evolutionary algorithms and artificial immune systems. In the last 20 years a lot of effort has gone into their theoretical foundation concentrating mainly on their efficiency, i.e., their expected performance, usually in context of optimisation. One common feature in many randomised search heuristics is mutation, an operator that takes as input a point in the search space and randomly generates a point in the search space as output. In artificial immune systems and many evolutionary algorithms it is the main exploration mechanism and understanding its features is of high interest. 

In this talk I will review some results on setting the mutation probability in evolutionary algorithms and contrast this with results about mutation in artificial immune systems which usually comes in the form of hypermutations. In addition to these concrete findings I will given an introduction to the modern analysis of randomised search heuristics.