Wonderful Life Utilities
- We can define three new WL utilities to experiment
with.
- In WLU0 we clamp down to "never attend". It is
the same as before: l(xdn(S)) -
l(xdn(S) -1)
- In WLU1 we clamp down to "always attend". It is SUMd != dn l(xd(S)) - l(xd(s)-1)
- In WLUa we clamp down to the "average action"
which in this case is attending with probability 1/h each
night.
- SUMd l(xd(S)) - SUMd != dn l(xd(S) + h/7) - l(xdn(S) - 1 + h/7)
José M. Vidal
.
11 of 13