This assignment explores using continuous optimization to perform discrete optimization. One example is generating grasps using learning (example1, Google Arm Farm reference, see also example3). Another is Contact-Invariant Optimization (video) ( Hands, Legs)
I would like you to reimplement one of these papers to see if their performance claims are correct. A simple version would be to reimplement Contact-Invariant Optimization, possibly simplifying the dynamics (we will discuss how in class).