¹¹To take another example: Tree search, which is essential planning, can also be parallelized, although it is a bit more difficult. After a few steps in the search tree, you can distribute the different branches to different processors, and each processor can search further in one of the branches. Intuitively, this would be like a boss assigning five different scenarios to her employees in a planning exercise. Each employee gets one scenario which each start from different assumptions, corresponding to the first branchings of the search tree. After the initial hurdle of formulating the scenarios (that is, building the initial branches of the tree), the parallelization is easy.