What's wrong with Multisim?

In post #6, i think it does not prove anything except that the mathematical calculatins are correct.
The results of Computer Simulation is on Matlab but are taken after what time when simulation started? It is not steady state. The charge on capacitors are rising initially with time.
I simulated and found that the result on start on MULTISIM is exactly the same as mathematical calculations done. I have posted screen capture at the start of simulation earlier .
It reaches the values of the matlab results after few minutes of simulation on my computer.
Only one multimeter was used practically. That is voltage was first checked on one capacter, and then on the other. Meter internal resistance was put in parallel to only one capacitor at a time. This condition is different than in schematic in simulator. Charge changes on capacitor with time. Sampling time is different and hence another reason for error. Sampling time of first reading is different than the sampling time of second reading as meter is used to take one sample and then taking second reading ( after noting it down).
For this test, voltmeters with sample and hold function simultaneously triggered would give better results.

An additional comment on the internal resistance point. "Virtual" MultiSim instruments have an internal resistance similar to real instruments. For the voltmeter used in the original post, the default value is as low as 1 Mohm. But it can be programmed to different values.

Another point is the minimal node conductance of SPICE simulators. The default value is 1e-12 (eq. 1 Teraohm) for most simulators. If no virtual instruments with finite resistance or other resistors are connected, the minimal node resistance will rule the steady state voltage of purely capacitive nodes, together with other simulation parameters, In this case, the results must not necessarily be congruent with Matlab or a hand calculation.