Matchmaking in p2p e-marketplaces: soft constraints and compromise matching

Automated matchmaking is increasingly used in e-marketplaces. A participant is presented with a ranked list of coun- terpart profiles of other participants. A participant profile consists of a set of hard (inflexible) and soft (flexible) con- straints. For two participant profiles, corresponding pairs of constraints are compared. We propose compromise match- ing wherein one or both constraints of a pair are soft and their values do not match exactly. We demonstrate com- promise matching with a home rental use case with data obtained from an existing e-marketplace.