A mathematical model for learning agents on a multi-agent system | IEEE Conference Publication | IEEE Xplore