Parameter Fitting in Neutron Scattering

Area: Chemistry
Neutron scattering is an important research tool destined to understand the molecular dynamics of solids and liquids. Its purpose is to figure out precise details of the positions, motions, and magnetic properties of atoms by investigating neutrons colliding with atomic nuclei and measuring their change of momentum and energy. Experimental data obtained from neutron scattering experiments is compared to a theoretical model implemented as a computer program. The adjustment of model parameters to experimental data is carried out by solving an unconstrained minimization problem. The optimization algorithm makes use of the derivatives of the objective function. The forward mode of automatic differentiation is applied to accurately and efficiently compute the corresponding derivatives.
Related ADTool:
ADIFOR
Reference:
People involved:
Reference:
Christian H. Bischof, H. Martin Bücker, Dieter an Mey
A Case Study of Computational Differentiation Applied to Neutron Scattering
Automatic Differentiation of Algorithms: From Simulation to Optimization, Springer, 2002
H. Martin Bücker, A. Rasch
Efficient Derivative Computations in Neutron Scattering via Interface Contraction
Conference proceeding, Proceedings of the 17th ACM Symposium on Applied Computing, Madrid, Spain, March 1014, 2002, ACM Press, 2002
A Case Study of Computational Differentiation Applied to Neutron Scattering
Automatic Differentiation of Algorithms: From Simulation to Optimization, Springer, 2002
H. Martin Bücker, A. Rasch
Efficient Derivative Computations in Neutron Scattering via Interface Contraction
Conference proceeding, Proceedings of the 17th ACM Symposium on Applied Computing, Madrid, Spain, March 1014, 2002, ACM Press, 2002
People involved:
 C.H. Bischof, H.M. Bücker, D. an Mey, A. Rasch,
Institute for Scientific Computing, RWTH Aachen University  J. Ette,
Institute of Physical Chemistry, RWTH Aachen University