Newton's method for fractal approximation (Q1120766)

The author solved the problem of finding the best approximation possible in the \(L^ q\) norm to an object by a fractal function whose graph is the attractor for an iterated function system dependent on several parameters. His approach is to define a potential function \(V(w)=\int_{D}v(x,w)-u(x)\|^ qd\mu\) of the parameters \(w=w_ 1,...,w_ k\) for the system, based on the \(L^ q\) norm of the difference between the target function \(u(x),x=(x_ 1,...,x_ n)\) and the function v(x,w) generated by the iterated function system. Then, he sought a minimum for the potential function corresponding to the superstable fixed point \(w^*\), by applying Newton's method to the gradient of the potential. The key component of the method is a technique for calculating the derivative of the potential with respect to the parameters w of the system. The author thus offers an efficient way to fit a target function with a function generated by an iterated function system.