multiagent_opt is a stochastic mixed-integer constrained optimization method based on multiagent learning. 

The repository is set up for the quadrotor optimization problem at https://github.com/hulsed/QuadrotorModel. 
* To demonstrate with the quadrotor problem, place that repository, along with Qprop.exe (http://web.mit.edu/drela/Public/web/qprop/) in a folder inside the optimizer directory. 
* To apply to different problem, place that problem in the directory, edit main.m and set the appropriate options in multiagent_opt.m

Further set-up instructions are shown as comments in main.m and multiagent_opt.m.

The repository requires Matlab's statistics toolbox.