- #1
wronski77
- 2
- 0
Dear forum members,
I am trying to solve the following system of equations.
ψ(x,y,z)=∫∫ψ(x',y',z)K(x',y',z)dx'dy'
z=f(ψ)
What I do is to solve the integral equation with a Monte Carlo method, evaluate "z" and do a loop until convergence.
My question to you is whether it is possible to accelerate the convergence by using stochastic approximation method such as Stochastic gradient descent (a.k.a. Robbins-Monro). I would highly appreciate any comments on the subject, including general information about the Robbins-Monro algorithm. What I know about the Robbins-Monro algorithm is that it is used to find zeros of nonlinear stochastic equations. Can It be applied to stochastic Fredholm problems like the one above?
Thank you in advance,
I am trying to solve the following system of equations.
ψ(x,y,z)=∫∫ψ(x',y',z)K(x',y',z)dx'dy'
z=f(ψ)
What I do is to solve the integral equation with a Monte Carlo method, evaluate "z" and do a loop until convergence.
My question to you is whether it is possible to accelerate the convergence by using stochastic approximation method such as Stochastic gradient descent (a.k.a. Robbins-Monro). I would highly appreciate any comments on the subject, including general information about the Robbins-Monro algorithm. What I know about the Robbins-Monro algorithm is that it is used to find zeros of nonlinear stochastic equations. Can It be applied to stochastic Fredholm problems like the one above?
Thank you in advance,