Mathematical optimisation often involves using the differential of the fitness function. In simple cases it may be possible to solve equations to find parameters such that the partial differential (slope) in every direction is zero. Alteratively, if the differentials can be calculated, one can use gradient descent to move in the direction of maximium improvement of fitness. Typically these techniques only work when the parameters are continuous variables the fitness function is continuous and differentiable – that is no sudden jumps or changes in slope.
Used in Chap. 4: page 51