Module 24: Learning Classifier Systems & Hyper-heuristics - Evolving Rules and Evolving Heuristics

LCS evolve rule-based systems that learn from environment feedback. Hyper-heuristics go a level up: instead of evolving solutions, evolve the heuristics that generate solutions. Both are powerful meta-level applications of EC.

Learning Objectives

1. Understand Michigan-style LCS (XCS) and Pittsburgh-style LCS
2. Implement a simple XCS for a classification task
3. Understand hyper-heuristics (selection vs generation)
4. Know GP-based hyper-heuristics
5. Understand automatic algorithm configuration (irace, SMAC)

Concept Explanation

Coming soon.

Code Examples

Coming soon.

Exercises

Coming soon.

Milestone Checklist

• Implemented a basic LCS
• Understand Michigan vs Pittsburgh distinction
• Know what hyper-heuristics are and when to use them
• Understand algorithm configuration vs algorithm selection