If you’re talking simple models, flat ground and constant speed then there won’t be any difference between 26, 29 or any other wheel size.
Correct. So that’s your starting model.
Next step would be to add in “number of accelerations”, some way of quantifying the time difference due to getting back up to your constant speed (at the start, after tight corners etc). That would tip things in the 26ers favour. For the basic model above, there are zero “accelerations” so 26=29.
After that you might add a “rollability” factor, i.e. the gain for a larger wheel over a given “roughness of terrain”. Again for a smooth course 26=29, as roughness increases so does the 29er advantage.
That would get you a reasonable way towards modelling a twisty, rough course at a constant altitude. If you wanted to try and track weight, you’d need to add in a height variance over the course distance.
It’s definitely do-able, but would get complicated quite quickly. Also, it would be very difficult to come up with the equation constants (e.g. by how much a given 90 degree corner slows down a 29er vs a 26er). Without a real-world testing ground (unlikely) it would be quite hard I think…