Idea 1 : Maze Generation

I would like to examine maze generation algorithms for the purpose of generating more challenging domains for search algorithms to solve. Creating domains that challenge existing search algorithms can assist in the development of more robust search algorithms that can avoid certain pitfalls of existing algorithms. Additionally, mazes are widely understood and have efficient state changes, which can allow for more algorithm based examinations in the future. I would like to develop a system for rating the “hardness” of a given maze, as well as creating a maze generation algorithm that can generate mazes that have a higher or lower “hardness” rating. 

Idea 2 : Cave Generation Using Cellular Automata

I would like to experiment with using cellular automata to generate cave structures. This has applications in procedural level generation for video games, artistic potential, and depending on the techniques used, it could also be useful in real world geological simulations. There has already been some work done in the area which gives ample room for extension and exploration. Most of the materials seem to be focused on 2D maze generation, so it may be fruitful to focus on generating 3D structures, or extending the existing techniques from 2D CA to 3D. 

Idea 3 : Origami Crease Pattern Generation Tool

I would like to build an application that generates an origami crease pattern from a 3D model, with a stretch goal being to implement 3D scanning from a camera. Folding techniques used in origami are seeing more and more uses in engineering contexts. The new Webb telescope for instance used origami for its heat shield and mirrors. Being able to generate crease patterns based on a source model may allow for more widespread use of similar techniques. This project will involve image analysis and potentially machine learning.