Optimization - Evaluation of cost function requires the cost functions value at the last iteration

Hi everybody,

I have an (unconstrained) optimization problem. In order to evaluate the function I am trying to minimize at iteration n (it maps from the complex plane to the real numbers), I need to know the location the optimization algorithm evaluated the function at iteration n-1. For example: I choose the starting point [0 0], the algorithm determines [1 1] as the next point. Then I need the point [0 0] as an input to the cost function to evaluate it at [1 1].

The examples provided in the MATLAB help seem only to deal with cases, where the cost functions value does not depend on any prior points. It thought about declaring a global variable for the points in the complex plane, but is this a good way to proceed? Are there other possibilities?

It would be very much appreciated, if anybody could give me a hint.

Thanks

Answer 1

[Guy Rouleau]

As far as I know, the cost function of any optimization problem should not depend on previous iterations.