Add implicit solver for long-time problems

[abisner]

An implicit integration method is needed in addition to the existing velocity-Verlet explicit integrator to make long-time quasi-static BVPs and/or steady-state problems tractable.

[yaowwwwwww]

Except for the implicit solver, add the ADR damping term into the explicit solver for rapid convergence is also a choice.

[JBludau]

The linked PR adds ADR integration. If any of you wants to try, feedback is appreciated. It can be seamlessly mixed with the Velocity-Verlet integration, see the added example case