How does local beam search utilise parallel search methods?

Select one:

By running independent searches from k different starting states without interaction.

By synchronizing kk searches to converge at the same solution.

By distributing the kk states across multiple processors to speed up the search.

By maintaining k iterations in parallel, each optimising a different part of the search space.

By exploring k different paths in the search space simultaneously and combining their results.