Originally Posted by

**bobb** hi captainblack,

1. i have a set of 8 data pairs, (x1, y1)...(x8,y8). i have to fit a model to these data. the actual model is very complicated, so i posted a sample model function, which is y(x) = A exp (x^2) + B cos(Cx) + Dx^0.4. there is only 1 dimension, and that is x; but there are 4 model parameters that i have to vary and find the values that give the best fit to the data points. my question is, is this considered a single-dimension or multidimension problem? i cannot find a clear-cut distinction between multidimensional and multivariable.

2. my reason for asking this question is that i would like to get a conceptual understanding of how this algorithm works. how is it that a problem of determining the global extremum of a function can be solved by defining a geometrical shape, and then manipulating that simplex to home in on the extremum? a short example, if you could, would be really great!