Look at the picture.
The figure on the left shows the individual box colliders in green.
The yellow boxes on the right are the results of the optimization of the original box colliders.
What is the algorithm that makes this possible?
Sure you can.
Assuming you have a list of
BoxCollider that need to be combined, you can access each of their
bounds properties, which consist of the boxes'
extents. With this two information in mind, you can easily calculate the minimum and maximum x, y, and z coordinates of the boxes.
All you need to do later is remove or disable initial
BoxColliders in your list.