I want to use latin hypercube sampling followed by Differential Evolution optimization algorithm in python to optimize material parameters for short fiber reinforced thermoplastics.
Dear Meghana, I have been working with parameter optimization techniques for more than 30 years. Why did you choose Nelder-Mead? Basically, the choice of optimization methods depends extremely on the formulation of the problem, the properties of the objective function, the existence of constraints, how these are integrated in the model functions etc. To give a reliable answer, more information is needed.
I am working on optimization of parameter estimation for deformation and failure prediction of SFRP (Short firber reinforced plastic). Nelder mead was chosen by the person who previously worked on my task, but as it is better for local optimization I decided to perform Latin hypercube sampling to get better idea of design space.
But recently I have moved from Nelder-Mead to Differential evolution as I thought it would be more suitable for my problem. Currently I am performing optimization with LHS + Differential evolution. I am not sure how this combination will result into. I would like to know if you have any idea about this.