Trajectory Optimization API

[alberthli]

This PR adds a very generic and flexible API for trajectory optimization. As a specific instantiation, it also implements the extremely simple predictive sampling algorithm.

• defines a generic API for trajectory optimization
• defines a generic API for cost functions (and their gradients) along with the ability to pass a generic set of caches/parameters to relevant functions
  • implements a simple quadratic cost function with fixed goal state
• implements predictive sampling

[alberthli]

Agreed on the first two requested tests - those should be trivial to spin up.

Sanity check of choice: randomly sample a batch of initial policies and also shoot them forward. pass those guesses to vanilla predictive sampling, which returns new trajectories that should be no worse than the guesses. verify this property in a test.

TODO list:

• [x] refactor q and v to x everywhere
• [x] test cost function and its derivatives (thanks for suggesting this, literally all my implementations were wrong)
• [x] smoke test for VanillaPredictiveSampler + optimize + jit
• [x] sanity check for VanillaPredictiveSampler
• [x] drop example file from PR after all other tests are implemented