Worst-Case Propagation of Uncertainty

worstcase attempts to maximize and minimize a function over give uncertainty bounds using fmincon. This function is superior to my old propagation of uncertainty code, PropError. Only the Optimization toolbox is required. Input is far easier, and implicit and non-analytic functions can also be handled. Execution time is significantly faster.
Two examples are given:
worstcase_example quickly demonstrates the method for a function.
worstcase_colebrook_example demonstrates error propagation for an implicit function. In the example, 25% uncertainty in the paramters of the implicit Colebrook equation.
Just run the two examples, and check out the code.