Prerequisites

• Lecture: "Natural and Artificial Evolution"
• Lecture: "Evolutionary Robotics"

Overview

The goal of this laboratory session is to deepen your understanding of evolutionary algorithms. After this session, you will (1) know how the morphology and the control of robots and artificial creatures can be co-evolved, (2) be familiar with an example of a genetic encoding for this purpose, and (3) understand basic strategies for setting the fitness function for multi-objective problems.

Note that the issues regarding the genetic encoding and the fitness function discussed here are not specific to co-evolution of morphology and control of robots, but reappear in similar forms in many evolutionary optimization problems.

Resources

• Framsticks GUI installer
• Framsticks manual
• Useful article by Framsticks creator M. Komosinski
• Framsticks homepage

Exercise 1: Genetic encoding and genetic operators

A detailed description of the exercise can downloaded here.

Exercise 2: Evolving creatures

A detailed description of the exercise can downloaded here.