todo.txt

# Jun 28
1. Fix E-grid_collect [X]
2. Critic update - check for further bugs (working commit) [X]
--- End Critic ---

3. Target Actor update - load actor problem not critic [X]
4. Local Actor update - follow similar format as Critic update [x]
--- End Actor ---

5. Make sure everything runs properly [X]
--- End TD3 ---

6. Generate plot w.r.t RBC, Optim (last week), and TD3 [] @Vanshaj



### ISSUES:

1. Clarification on reward warping --> optimization (E_grid is a var) or taking data from Actor.forward? ### Solve it again, correct.
2. If optimization, objective is maximized? (#L330)
3. If optimization, peak_net_electricity_cost (#L201) square DCP violation. Using cp.norm(x, 2) still causes issues.
4. Running update for alphas in critic update per day within buffer.sample()? (see TD3.py)
5. Critic update ---> primal infeasable debugging. (19th hour of 2nd day of meta-episode) (see debug.ipynb)
6. Actor update --> reward_warping_loss (sum or mean)?
7. Actor update --> grad is of dimension 9 (buildings), taking the mean across that