(Text enbart på engelska)
The ambition is to cover a large part of the optimization field, including linear, non-linear, integer, and infinite dimensional problems. We work over a broad spectrum in optimization; theory, algorithms, and most important the development of tools and software in collaboration with optimizatioon experts at other universities and with experts in natural science.
We have developed software that have been successfully used in diverse areas such as Radio Stereometric Analysis, the training phase of Large feedforward neural networks and some ODE applications.
The current research is focused on the computation and analysis of approximate solutions of non-linear systems of equations. We develope algorithms that, in an efficient and flexible way, can handle such issuses as ill-conditioning, sparsity, constraints, weights, and perturbed solutions.
