Now that continuation can return initial guesses, one strategy I think would be helpful is a node pruning strategy. It doesn't change the mission, but rather tries to solve the same mission with less nodes. This is useful because, sometimes an initial guess or continuation may use a huge number of nodes, then later a lot of nodes aren't required but are included anyways. This slows down bvp solvers that use every node in calculation.
Doing this manually, I've gotten ~90% speedup for some problems.
Now that continuation can return initial guesses, one strategy I think would be helpful is a node pruning strategy. It doesn't change the mission, but rather tries to solve the same mission with less nodes. This is useful because, sometimes an initial guess or continuation may use a huge number of nodes, then later a lot of nodes aren't required but are included anyways. This slows down bvp solvers that use every node in calculation.
Doing this manually, I've gotten ~90% speedup for some problems.