In today’s highly interconnected world, many situations involve a large number of agents taking decisions. Mean field games have been introduced to cope with the fact that individual interactions become intractable as the size of the population grows. The key idea is to replace the microscopic viewpoint by a macroscopic one, describing the interactions of a typical player with the population’s distribution. Although such a mean field approximation leads to considerable simplifications, computing solutions, which is of crucial importance for applications, remains a challenging task. In this talk, we present several numerical methods and illustrate them on typical examples. Borrowing tools from stochastic analysis, optimization and machine learning, these methods enable us to compute the evolution of the population's distribution as well as the optimal behavior.