Leader Selection Method of Leader Oriented Matchmaking Algorithm

Leader-Oriented Matchmaking (LOM) is a matchmaking algorithm that selects leaders and member candidates from the players and pairs them using a minimum-cost maximum-flow algorithm. Unfortunately, LOM only states random selection to select the leaders. LOM needs a sure way to select the leaders that ensures it always produces good, consistent results since different leader combinations could produce different results. Thus, this research proposes three leader selection methods for LOM: the Largest-Gap, Room-Capacity, and N-Max-Gap. We tested all three methods along with the random selection method stated in LOM's paper to find the best leader selection method. We analyze the closeness between leaders and their members and the closeness between players. As a result, if we prefer the high closeness of the players, the Largest-Gap method is the best to use since it produces the lowest average association weight between players (24.58% better than average result) with the highest consistency. Then, if we prefer the high closeness of the leaders and their members, the Room-Capacity method is the best since it produces the lowest average association weight between leader and its members (49.68% better than average result) with the highest consistency. An association weight indicates the closeness between two players, whereas a lower association weight means the closer the properties between two players are.