Evolving Solutions to Computational Problems Using Carbon Nanotubes
Maktuba Mohid and J. F. Miller
Evolution-in-materio is a method that uses artificial evolution to exploit properties of materials to solve computational problems without requiring a detailed understanding of such properties. In this paper, we show that using a purpose-built hardware platform called Mecobo, it is possible to evolve voltages and signals applied to physical materials to solve two computational problems: classification and robot control. We have investigated many experimental factors that could affect the effectiveness the device and the evolution-in-materio technique. On the classification problem (Iris) we show that the evolution-in-materio approach with analogue signals gives very good results that are competitive with well-known effective software-based evolutionary approach. In the case of robot control, we were able to evolve a controller that successfully allowed a simulated Khepera robot to fully explore its environment without colliding with any obstacle.
Keywords: Evolutionary algorithms evolution-in-materio material computation evolvable hardware machine learning classification robot control