"I was very pleased to receive the award. The news reached me in Finland, where I am working with the same colleagues who acted as consultants during my master's thesis – naturally, we celebrated it properly," says Daniel Reeves.
The melting of glaciers and ice sheets poses a significant threat to ecosystems, making accurate modelling of their thermodynamic behaviour an important area of research. Such simulations also involve problems with inequality constraints, which can be formulated as quadratic programming (QP) problems. The Modified Proportioning with Reduced Gradient Projections (MPRGP) algorithm provides an efficient approach to solving these problems.

In his thesis, Daniel focused on applying this algorithm to problems in glaciology, specifically to solving the heat conduction equation with a simple upper inequality constraint on temperature. The work also involved developing an interface between the PERMON library and Elmer/Ice, a software tool designed for glacier modelling, and applying methods from the PERMON library to solve QP problems.

The proposed solution was subsequently tested on a coupled thermomechanical steady-state simulation of the Midtre Lovénbreen glacier in Svalbard. The results also demonstrate the potential to improve the scalability of solvers for glacier melting simulations, which could enable larger-scale simulations, such as modelling entire Greenland or Antarctica, in the future.

"Daniel is an excellent student who produced a highly above-average master's thesis. He studied the new field of glacier melting and applied quadratic programming algorithms to it, which he implemented himself. Moreover, the results improving the scalability of glacier-melting solvers make it possible to perform larger-scale simulations, such as of entire Greenland and Antarctica, and could serve as a basis for upcoming European projects," says the thesis supervisor, David Horák.
Daniel Reeves will continue building on the results of his master's thesis in his doctoral studies at VŠB–TUO.

"I would like to continue with doctoral studies at VŠB, where I will focus on the research and implementation of quadratic programming algorithms for real-time motion planning in automated mobility," he says.