Hrothgar is a parallel minimizer and Markov Chain Monte Carlo generator by Andisheh Mahdavi of San Francisco State University. It has been used to solve optimization problems in astrophysics (galaxy cluster mass profiles) as well as in experimental particle physics (hadronic tau decays). It is probably adaptable enough to be applied to your merit function if you can write it in C.
- Link with any C or compatible language merit function
- Studies your merit function and calculates a sampling covariance matrix on the go
- Run Levenberg-Marquardt downhill fits or full MCMC---your choice
- Multithreaded graphic visualization with PGPLOT---watch your chains grow on the screen
- Run chains under MPI and a master process will monitor them for convergence
- User OpenMP in your merit function without disrupting the MCMC
Be the first to post a review of Hrothgar Parallel LM/MCMC Minimizer!