Prof. Dr. Malte Helmert was arwarded a SNSF Advanced Grant
Towards a unified theory of state-space search
In Artificial Intelligence, abstractly representing a problem using a state space and searching for valid paths is one of the most important ways to solve a problem. However, finding paths in very large graphs is a central problem in AI and other areas of computer science, since graphs of extreme size are common to many combinatorial optimization problems.
In AI research, state-space search is referred to as automated planning. It has been intensively researched for decades and has produced three predominant techniques, all of which have their specific strengths and significantly outperform the other two approaches on some inputs.
The project of computer scientist Malte Helmert now aims to develop a unifying theory that can explain the mechanisms by which each approach achieves selective superiority. In a second step, the researchers want to develop novel algorithms that combine the best features and can dominate existing algorithms from different methods.
In this way, the project promises to provide a new, unified understanding of the algorithmic foundations of factorized state-space search and to bring together three currently unrelated research directions with enormous potential for synergy.
Malte Helmert has been Professor of Computer Science at the Department of Mathematics and Computer Science since 2011. In 2013, he was awarded a Starting Grant and in 2018 a Consolidator Grant by the European Research Council (ERC).