Prospects for Adaptation

  • Tim Kovacs
Part of the Distinguished Dissertations book series (DISTDISS)


This Chapter considers the conditions under which XCS and SB- XCS can be expected to adapt successfully, and in so doing develops some of the main results of the thesis. In particular, it develops a taxonomy of different rule types and considers the conditions under which they may occur. In order to do so an extreme simplification of the classifier system is made, which forces us towards qualitative rather than quantitative analysis. We begin with the basics, consider- ing definitions for correct and incorrect actions, and then correct, in- correct, and overgeneral rules for both strength and accuracy-based fitness. The concept of strong overgeneral rules, which we claim are the Achilles’ heel of strength-based classifier systems, are then anal- ysed. It is shown that they depend on the structure of the reward function (or in sequential tasks, the value function). We distinguish between strong and fit overgeneral rules, and show that although strong overgenerals are fit in SB-XCS, they are not in XCS. We show how to design fit overgeneral rules for XCS, and consider why we might want to use reward functions which may produce strong overgenerals. Finally, we return to the Woods2 task of Chapter 2 and explain SB-XCS’s performance on it using theoretical arguments and empirical evidence.


Correct Action Action Selection Rule Allocation Reward Function Sequential Task 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag London 2004

Authors and Affiliations

  • Tim Kovacs

There are no affiliations available

Personalised recommendations