Quiz: Local Search

Methods of AI — 30/04/26


Q1 — Local Beam Search vs. Parallel Hill Climbing

Question: Local Beam Search selects the k best successors across all neighbors of all k current states combined, while Parallel Hill Climbing runs k searches independently. Why does this sharing of information make Local Beam Search fundamentally different — and what failure mode does it introduce that Parallel Hill Climbing doesn’t have?

Max’s answer:
Result:


Q2 — Simulated Annealing

Question: Simulated Annealing can accept downhill moves. Explain the acceptance formula, what role the temperature T plays, and why a slow cooling schedule is necessary for finding the global optimum.

Max’s answer:
Result:


Q3 — Genetic Algorithms

Question: What is the uniform-order crossover, and why is standard 1-point crossover insufficient for permutation-encoded problems like TSP?

Max’s answer:
Result:


Score

Total: / 3