# A manifold of agents?

In contract theory and mechanism design, if the agents have one-dimensional characteristics, then they are usually considered as a contiuum on $\mathbb R$; if they have n-dimensional type, then they are in a contiuum on $\mathbb R^n$.

Naturally, if those (n-dimensional) types have some sort of correlations, then we could have a manifold of agents!

Or, if "n" is infinite, then we have agents on the infinite dimensional vector space.

Are there any research considered a manifold of agents?

• Something like services.bepress.com/wilson/art20 ? – Michael Greinecker Oct 10 '17 at 15:12
• @MichaelGreinecker I was reading the same paper, too! This paper could be considered as "agents' types are on a infinite dimensional vector subspace $V$", and sometimes, a general topological manifold. – High GPA Oct 10 '17 at 19:30