Quiz on Local Search Algorithms and Online Search

What is the primary goal of local search algorithms?

• To explore all possible states

• To find the solution by minimizing or maximizing an objective function

• To maintain a search tree

• To track multiple states simultaneously

In hill-climbing search, what does the algorithm do when it reaches a peak?

• It explores all neighboring states

• It resets to the initial state

• It terminates as no neighbor has a higher value

• It continues to search for better states

What is a local maximum?

• A state that leads to a dead-end

• A state with the same value as all its neighbors

• A state that is better than its neighbors but not the best overall

• The best possible state in the state space

What is the main advantage of local search algorithms?

• They maintain a complete search tree

• They can find solutions in large or infinite state spaces

• They require a lot of memory

• They always find the global maximum

What is a plateau in the context of hill-climbing?

• A flat area where all neighbors have the same value

• A state that is unreachable

• A peak with no higher neighbors

• A state that leads to a global maximum

What does random-restart hill climbing involve?

• Using a single starting point

• Restarting the search from a random state when stuck

• Maintaining multiple states at once

• Always moving to the best neighbor

What is simulated annealing based on?

• A type of online search

• A process used in metallurgy

• A genetic algorithm technique

• A method of hill climbing

In genetic algorithms, what is the purpose of crossover?

• To evaluate fitness

• To select the best state

• To combine two parent states to create offspring

• To mutate a state

What is the competitive ratio in online search?

• The ratio of the number of states visited

• The ratio of successful actions to total actions

• The ratio of online path cost to best path cost

• The ratio of time taken to find a solution

What is an online search agent required to do?

• Avoid any form of exploration

• Maintain a complete search tree

• Perform actions to determine outcomes

• Know the entire problem data in advance