Which local search algorithm accepts bad moves with probability $P = \exp(delta_E / T)$ to escape local optima?
Simulated Annealing borrows from metallurgy, accepting worse moves with probability $\exp(delta_E / T)$ where T decreases over time.
In an AND-OR graph, which search algorithm is specifically designed to find optimal solution trees?
AO* search operates on AND-OR graphs using heuristic estimates to find optimal subtrees.
When does A* search terminate with an optimal solution?
Testing for goal upon generation does not guarantee optimality in A*; it must be popped from Open.
In Expectimax search, chance nodes evaluate their value as:
Chance nodes in Expectimax compute the probability-weighted sum of their successors.
What happens in Hill Climbing search when the algorithm reaches a PLATEAU?
On a plateau, neighboring states have the same evaluation, so there is no preferred uphill move; the search may stall or wander.