Mean field type control problems and mean field games can be viewed as models for strategic decision making in very large populations. In this talk, we will first explain the differences between these two theories, both in terms of motivation and modeling. Then, we will see how a dynamic programming principle can be proven for mean field type control problems. To do so, starting from a stochastic formulation we rewrite such problems as the optimal control of a McKean-Vlasov equation. We can also clarify a link with the calculus of variations for these problems. Last, a fixed point algorithm will be presented together with numerical results. This is joint work with Olivier Pironneau. Speaker(s): Matthieu Lauriere (NYU Shangai)