Skip to main content

Module 11: Estimation of Distribution Algorithms - From Operators to Probabilistic Models

What if instead of crossover and mutation, you build an explicit probability model of good solutions and sample from it? EDAs replace genetic operators with statistical modeling, and CMA-ES is secretly one of them.

Learning Objectives​

  1. Understand the EDA paradigm (model, sample, select, update)
  2. Implement PBIL and UMDA for binary problems
  3. Implement the compact GA (cGA)
  4. Understand BOA (Bayesian Optimization Algorithm)
  5. See CMA-ES through the EDA lens

Concept Explanation​

Coming soon.

Code Examples​

Coming soon.

Exercises​

Coming soon.

Milestone Checklist​

  • Implemented PBIL and UMDA
  • Understand the connection between EDAs and CMA-ES
  • Can explain when EDAs outperform operator-based EAs
  • Implemented cGA

Was this page helpful?