Latocha, D; Arabas, P
The task of choosing players to form optimal games is formulated as a binary mathematical programing problem. An efficient heuristic is proposed and decomposed to build a multi-agent system. The quality of such a solution is compared with the quality of the exact one, found with the MILP solver. The scalability of the system is verified by numerical experiments.
Keywords: Matchmaking MILP Multi-agent system