KKT Conditions for Dynamic Optimization

When would you use Lagrange multipliers during a direct transcription problem like this? They weren't used here. Whenever I think nonlinear optimization I think of Lagrange multipliers. Maybe I'm missing something.