2.3 Stochastic/Deterministic hybrid methods

 

Analysis, development and implementation of hybrid methods, combining Monte Carlo stochastic algorithms and deterministic discretization algorithms, for the solution of complex PDE problems.

 

The research activities of the action focus on the development of hybrid methods for the solution of complex PDE problems, which consist of a Monte Carlo stochastic process and deterministic methods (finite elements, finite differences). The stochastic process disintegrates the initial problem into a group of independent subproblems while the deterministic methods estimate solutions of the subproblems. We are confident that we could create both the appropriate framework for the solution of complex problems and a practical simulation tool.

The implementation of those schemes in modern computing environments is of great interest, not only because of the inherent parallelism stochastic methods present but also because of other additional features those schemes have, as far as their parallelism potential is concerned: small ratio of computations/communication, flexible flow-mechanisms, easy to implement on various computing standards (multithreading, cluster, web-services, etc).

 

Deliverables

 

  • 1 technical report
  • Publication of at least 3 scientific articles in international scientific journals or/and conference proceedings.
  • Software

 

The University of Thessaly Research Team is the main work group that will carry out the greater part of the action, will author and publish the results and edit a technical report with the description of the scientific activities that took place during the action and the research results that emerged.