Notes on deviating Tset from the following:

Tset = 
[  27.9042;
   27.3129;
   24.6523;
   19.4891; (1)  ---
   23.2008; (2)     |- (4)
   23.5146; (3)  ---
   19.9807;
   27.6923;
   25.2496;
   20.5860; (5)
   25.8963;
   27.0680;
   28.6026;
   21.7304; (6)
   23.0020;
   24.7590;
   24.4885;
   25.1275;
   26.4175;
   27.1974;
   33.7928;
   28.0392;
   27.6728;
   25.9199;
   26.9278];
Price to beat: $22.78

 Change notes
 -------------------------

 (1) Changing only this one from on -> off results in a cost increase of $0.2
 (2) On -> off gives cost increase of $0.6
 (3) Basically the same as (2)
 (4) Turning all of these from on -> off causes a net increase of $0.4 (keep only 1 pre-chilling hour).
 (5) Decreasing this to 20C (minus 0.5C) causes a price increase of about $0.038.
 (6) Changing this to 20C causes a minor SAVINGS of $0.1 in cost. However decreasing it further to 19C causes
	 the price to increase by ~$0.03 and so forth. 
 
 Price schedule notes
  -------------------------
[date, cost] = importLambda();   % Import lambda for price data
for i = 1:365
	stairs(cost(i,:))
	drawnow
	title(['Day = ', num2str(i)])
	pause
end
 32 (a lot like 257)
 50 (typical)
 66 (high evening prices that steadily increase over the day)
 67 (more extreme case of the above)
 68 (no extreme dips, but there are long periods where cooling is more affordable)
 69 (sinusoidal behavior but overall price doesn't vary that much)
 76 (wow, like 68 but if there WERE extreme dips)
 134 (I think this one might be pretty realistic)
 215 (looks like cosine rather than sine)
 222 (likely heavy pre-chilling strategy)
 236 (heavy evening costs again; drastic increase)
 242 (reasonable "nearly constant" price case with cheap morning hour)
 257 (two dips in price - early morning and afternoon - that aren't super extreme)
 
 35.8385
 