Artificial Intelligence - Adversarial Search - a Game with uncertainty 1. In the following, a “max” tree consists only of max nodes, whereas an “expectimax” tree consists of a max node at the root...

Artificial Intelligence - Adversarial Search - a Game with uncertainty

1. In the following, a “max” tree consists only of max nodes, whereas an “expectimax” tree consists
of a max node at the root with alternating layers of chance and max nodes. At chance nodes, all
outcome probabilities are nonzero. The goal is to find the value of the root with a bounded-depth
search. For each of following statements, either give an example or explain why this is impossible.

a) Assuming that leaf values are finite but unbounded, is pruning (as in α-β pruning) ever possible
in a max tree?
b) Is pruning ever possible in an expectimax tree under the same conditions?
c) If leaf values are all nonnegative, is pruning ever possible in a max tree? Give an example, or
explain why not.
d) If leaf values are all nonnegative, is pruning ever possible in an expectimax tree? Give an
example, or explain why not.
e) If leaf values are all in the range [0,1], is pruning ever possible in a max tree? Give an example,
or explain why not.
f) If leaf values are all in the range [0,1], is pruning ever possible in an expectimax tree?
g) Consider the outcomes of a chance node in an expectimax tree. Which of the following evaluation
orders is most likely to yield pruning opportunities?
g.i) Lowest probability first.
g.ii) Highest probability first.
g.iii) Doesn’t make any difference.