Evolving Modular Fast-Weight Networks for Control

Cooperative Co-Evolving Neural Networks for Robosoccer Simulation

Evolving Symmetric and Modular Neural Network Controllers for Multilegged Robots

Controllers for multilegged robots are characterized by modularity and symmetry. However, the controller symmetries necessary for generating appropriate gaits are often difficult to determine analytically. This paper utilizes a nature-inspired approach called Evolution of Network Symmetry and mOdularity (ENSO) to evolve such controllers automatically. It uses group theory to mutate symmetry sys...

Spatially Constrained Networks and the Evolution of Modular Control Systems

This paper investigates the relationship between spatially embedded neural network models and modularity. It is hypothesised that spatial constraints lead to a greater chance of evolving modular structures. Firstly, this is tested in a minimally modular task/controller scenario. Spatial networks were shown to possess the ability to generate modular controllers which were not found in standard, ...

Evolving Reusable Neural Modules

Topology and Weight Evolving Artificial Neural Networks (TWEANNs) have been shown to be powerful in nonlinear optimization tasks such as double pole-balancing. However, if the input, output, or network structures are high dimensional, the search space may be too large to search efficiently. If the symmetries inherent in many large domains were correctly identified and used to break the problem ...

A Direct Power Feeding System for AC Railway Networks Using Modular Multilevel Converter

Abstract Traditional railway power supply systems impose substantial power quality problems (PQ) on the utility network, such as unbalance, harmonics and a large amount of reactive power. This paper proposes a topology based on three-phase to single-phase modular multilevel converters (MMC) to obviate these problems. The MMC based traction substations (TSS) are connected directly to the utili...