milestone results

potential problems of the RL algorithm

1. RL state does not know about physical state: trajectory loops potentially dangerous for convergence

2. Replays/Forced Learning induce overfitting: non-best (s,a) pairs also updated thru the tilings. This would not occur in a tabular algorithm but this is also the main reason why tabular methods learn slower.

RL Data on the 2LS model

bang-bang and continuous protocols. Studied fidelity, energy increase above inst energy, energy fluctuations, diagonal entropy (basis of final state) and the BLoch sphere evolution of states.