Running Simulations

Initial Conditions

• The distribution function
• Example: The Hernquist model (from Hernquist 1990, ApJ, 356, 359)
• Monte Carlo sampling a distribution function
• Initializing an exponential disk -- radial distribution
• initializing a uniform disk -- the bad thing that happens when you sample evenly in r
• initializing a uniform disk -- the good thing that happens when you sample evenly in r² (i.e., area)
• monte carlo rejecting to make an exponential disk
• the sm code that did this
  
• Initializing an exponential disk -- thickness
• Initializing velocity dispersions

Running the Simulation

• Integrating the orbit -- the simplest (but not best) way
• Leapfrog integrators
• Runge-Kutta integrators
• Choosing a timestep I
• Choosing a timestep II
• Choosing a timestep III

Quality Control