Nicolas Greber wins ACCL Master Award

In his Master's thesis, Nicolas Greber examined the behaviour of self-learning pricing algorithms. Recent studies have shown that certain AI pricing algorithms are able to learn tacit collusion in oligopoly models. Nicolas picked up there with his master's thesis, analysing the behaviour of two more recent AI pricing algorithms. The first algorithm, known as the "deep deterministic policy gradient", failed to achieve constant convergence. In contrast, the second algorithm, the "multi-agent deep deterministic policy gradient", achieved consistent results in a Nash equilibrium. These findings were confirmed by various robustness tests.

We wish Nicolas all the best and success in his PhD at the University of Zürich and in his work for Swiss Economics.