Step 1 - Generative Design Framework
Step 2 - Generative Design Study
The outputs are computed in the Dynamo workspace as shown below:
Step 3 - Generative Design Study Results
The Dynamo study script is as follows, going left to right. Underneath is a zoom-in of the formation of the envelope cube, the shaft cube, and the floor surfaces that make up the building mass.
After running the optimization (10 generations of 20 samples), I found that the addition of basement floors definitively increased the elevator round trip time (because the passengers have to wait for the elevator to cycle both above and below ground). This was the opposite of what I expected. It is dependent, though, on the RTT modeling assumptions (no passengers on the return trip) and my choice to use all elevators in both directions.
So, the optimal choice was to place the bottommost floor at ground level. The optimization reduced down to a tradeoff between occupiable floor space and the elevator round trip time.
The floor area decreases linearly with the number of shafts. But the round trip time decreases nonlinearly. The largest time reduction was made going from 1 elevator/stair core to 2 cores (bolded in blue below). For this reason, I interpret the 2-core option as the optimal design.
