Method for maximizing standard deviation (maximize instantaneous uncertainty) in matlab using optimization

I am writing a code used to maximize the standard deviation of some design variable. Basically I want to maximize the uncertainty from one instant to the next in its operation. What is the methodology for doing this in general with matlab?

I believe I will need to use a combination of a linear regression model or a radial basis function, and then use fmincon or some other optimization function in matlab to parameterized my objective function. Any help appreciated!