Over some portion of the range of variables that you permitted, your equations are steep enough that double precision numbers are not sufficient to be able to find a reasonably smooth linear approximation between adjacent representable numbers. For example, x^10 is unreasonably large by the point of x = 40
- there might be a problem in the equations
- you might have to restrict the range of variables
- you might need to adjust the options to permit a larger tolerance
- you might have to switch to the symbolic toolbox.
