14. Interactive Evolutionary Algorithms

This chapter discusses the topic of interactive evolution, where the measure of a solution’s fitness is provided by a human’s subjective judgement, rather than by some predefined model of a problem. Of course, the world around us is full of examples of human intervention in biological evolution, in the form of pets, garden flowers, food crops and farm animals. Applications of Interactive Evolutionary Algorithms (IEAs) range from capturing aesthetics in art and design, to the personalisation of artefacts such as medical devices.

When including humans `in the loop’ we must consider their peculiar characteristics. On one hand, they can provide insight and guidance beyond simply selecting parents for breeding. On the other, they can be inconsistent, and are prone to fatigue and loss of attention. These factors make it inappropriate to use the `traditional’ model of an EA generating possibly thousands of candidate solutions. This chapter describes and explains some of the major algorithmic changes that have been proposed to cope with these issues.

Contents:
14.1 Characteristics of Interactive Evolution…………………215
14.1.1 The Effect of Time …………………………..216
14.1.2 The Effect of Context: What Has Gone Before . . . . . . . . 216
14.1.3 Advantages of IEAs…………………………..217
14.2 Algorithmic Approaches to the Challenges of IEAs . . . . . . . . . . . 217
14.2.1 Interactive Selection and Population Size . . . . . . . . . . . . . 217
14.2.2 Interaction in the Variation Process………………218
14.2.3 Methods for Reducing the Frequency of User Interactions218
14.3 Interactive Evolution as Design vs. Optimisation . . . . . . . . . . . . 219
14.4 Example Application: Automatic Elicitation of User Preferences220

Suggested Reading

Exercises

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

The on-line accompaniment to the book Introduction to Evolutionary Computing