Measuring the Complexity of ANN-produced Evolutionary Designs

There is promise in the field of Evolutionary Design for systems that evolve not only what to manufacture but also how to manufacture it. EvoFab is a system that uses Genetic Algorithms to evolve Artificial Neural Networks (ANNs) which control a modified 3d-printer with the goal of automating some level of invention. ANNs are an obvious choice for use with a system like this as they are canonically evolvable encodings, and have been successfully used as evolved control systems in Evolutionary Robotics. However, there is little known about how the structural characteristics of an ANN affect the shapes that can be produced when that ANN controls a system like a 3d-printer. We consider the relationship between certain structural characteristics of an ANN and the ability of that ANN to produce complex geometric shapes by controlling a 3d-printer. We develop an understanding of shape complexity for 2d-shapes in a simulated 3d-printer in order to use Genetic Algorithms to optimize ANNs with fixed structures to produce “complex” outputs and assess the relationship between topologies of ANNs and the system’s success in producing complex outputs.