Learning efficient satisfaction equilibria via trial and error | IEEE Conference Publication | IEEE Xplore