Unformatted text preview: ). n 5. (20+5 pts.) Search and MDPs (a) (3) The following diagram represents the state space of a deterministic problem, with each arrow denoting a possible operator (labelled with the step cost). Assume that the successors of a state are generated in alphabetical order, and that there is no repeatedstate checking. Show the search tree generated by breadth rst search applied to the problem of starting in A, where C is the goal. Circle the tree node that the search identi es as the solution.
B 1 A 6 C 4 (b) (2) Of the four algorithms breadth rst, depth rst, iterativedeepening, and uniformcost search, which nd a solution in this case? (c) (2) Of the four algorithms, which nd an optimal solution in this case? (d) (3) In MDPs, the values of states are related by the following equation, the Bellman equation (p.503): a U(i) = R(i) + max Mij U(j) a X
j a where R(i) is the reward associated with being in state i and Mij is the probability of reaching state j if action a is executed in state i. Su...
View
Full Document
 Spring '08
 Staff
 Computer Science, Logic, Artificial neural network, neural network, decision problem, C1 j= C2

Click to edit the document details