Standard procedure:
Pick a random Realization at high Z => run forward => see what you got
Constrained Simulations:
 | Start with the real data at z=0 |
| Get the large-scale density field |
| Extrapolate it back to a high redshift: Use linear theory and some cosmological model |
| Use the the density field as a constraint for setting random a realization for given cosmological model |
| Add short-scale waves |
| Mark III catalog of distances and velocities for galaxies |
| Wiener filter reconstruction on large scales (~5Mpc, but variable) (Zarobi et al. 1995) |
| Constrained realizations for 80Mpc/h radius (Hoffman & Ribak, 1991) |
| LCDM standard model: Omega_0 =0.3 Omega_L =0.7 h=0.7 |
| Use linear theory to go back to high redshifts |
| Add small scale waves where appropriate |
| Run simulations forward |