Yes, you can simulate a 7-post shaker on a computer, but like with anything in the world of engineering there is a trade-off. To get the most accurate results, you need to use nonlinear models, which are difficult to setup and take a lot of time that some teams do not have. You can linearize the models, but you limit the range of data that you can examine accurately depending on the point in the data about which the linearization is made, and the engineers that are developing the simulations are aware of this. Even if a team does have the capability to develop these models (I know Ferrari has used a computer simulation of a 7-post shaker), they still need to validate. I know this has become a bit of a buzzword around here in the aerodynamics threads, but it is very important. If an engineer thinks that they fully understand every minute element of an evolving system as complex as an F1 car, they are being naive and overestimating themselves, period.
So, let's set the record straight on what is being accomplished with the 7-, or 8-, post rig. They are not replacing track testing, they are supplementing track testing. Additionally, they are not replacing computer simulation, they are supplementing computer simulation. Two areas of the suspension that have a great effect on handling and response are the tires and dampers. Constant efforts go into modeling the effects of variables like temperature, pressure, slip angle, normal load, etc on the lateral and longitudinal forces developed at the contact patch in addition to all of the moments that are produced. It is straightforward to use this tire data to develop a steady-state cornering simulation based on the springs and suspension geometry being used. The variable that is missing is the damper, and this is where the 7-post shaker rig is useful. Dampers are primarily used as control elements. They are used to control the variation in normal load, and thus the variation in lateral and longitudinal force at the contact patch. Additionally, they are used to control pitch and heave of the chassis, which is critical on an aero-dependent car.
The shaker rig is not always used to replicate track data. Engineers can set a plaid of damper settings between the front and rear dampers and see what effect they have on the heave and pitch response resulting from a controlled input. The input can be a swept sine curve or it can be noise with some dominant frequency. Generally, you can either minimize the pitch, or minimize the heave, but you cannot minimize both (see the second paper that I linked previously... "Damper tuning with the use of a Seven Post Shaker Rig"). The setup engineers can then look at this data and determine the "best baseline" setup. In fact, in many cases doing this has produced quicker lap times than driver feedback. The drivers still have some input, though. They can take the car in the baseline setup, and can provide feedback on how controllable the car is in a given track situation. Based on this driver feedback and telemetry data, the engineers can then look at data from the shaker rig and determine how much they may want to change the damper settings, and not just a general direction of the change that needs to be made. This way they are engineering and not tweaking the setup. Doing this minimizes the track time required, which is very important considering the limitations imposed on track testing.
The bottom line is that computer simulation, controlled lab testing, and track testing are not mutually exclusive. They all complement each other, and the whole is greater than the sum of the parts.