Consider the following: (a) Trapping at local maxima (b) Reaching a plateau…

Answer: All three listed problems — trapping at local maxima, reaching a plateau, and traversal along the ridge — are shortcomings of the hill climbing algorithm.

• Trapping at local maxima: Hill climbing is greedy and only accepts improvements, so it can converge to a nearby peak that is not the global optimum and cannot escape.

• Reaching a plateau: When many neighbors have equal evaluation, there is no improving move and the algorithm stalls or wanders without progress.

• Traversal along the ridge: Narrow ridges may require moves that temporarily decrease the objective or sideways moves to continue toward the global optimum; basic hill climbing does not perform such moves effectively.

Mitigations:

• Random restarts: run the algorithm from multiple starting points to reduce the chance of all runs getting stuck in poor local optima.

• Allow sideways moves or limited non-improving moves: helps cross plateaus or move along ridges.

• Use stochastic variants or global methods (e.g., simulated annealing, genetic algorithms) when the search space has many local optima.