matteobutano Goto Github PK

backward_forward_solver

We use MATLAB to solve a backward-forward system of coupled Mean-Field Games Differential equations.

mfg_ergodic_intruder

Where we solve the NLS version of a Mean-Field Game simulating the passage of a cylindrical intruder through a static dense crowd.

mfg_pinns

neural_mfg

We use Deep Galerking Method (DGM) to solve the Non-Linear Schrödinger Formulation (NLSF) of a Mean-Field Game describing the passage of a cylindrical intruder through a dense crowd.

optimal_crowds

Use optimal control theory to simulate the dynamics of crowds. Set up a room, initialise agents and their targets. The solution of a Hamilton-Jacobi-Bellman equation provides the optimal trajectories towards the goals, while the actual motion is simulated using an accurate social force model.

proteins_ml

Where we use supervised and unsupervised learning to analyse a large database of proteins.

time_series_ml

Where we use Neural Network to study and predict stock prices time series.