Towards Automated Bargaining in Electronic Markets: A Partially Two-Sided Competition Model

This paper focuses on the prominent issue of automating bargaining agents within electronic markets. Models of bargaining in literature deal with settings wherein there are only two agents and no model satisfactorily captures settings in which there is competition among buyers, being they more than one, and analogously among sellers. In this paper, we extend the principal bargaining protocol, i.e. the alternating-offers protocol, to capture bargaining in markets. The model we propose is such that, in presence of a unique buyer and a unique seller, agents' equilibrium strategies are those in the original protocol. Moreover, we game theoretically study the considered game providing the following results: in presence of one-sided competition (more buyers and one seller or vice versa) we provide agents' equilibrium strategies for all the values of the parameters, in presence of two-sided competition (more buyers and more sellers) we provide an algorithm that produce agents' equilibrium strategies for a large set of the parameters and we experimentally evaluate its effectiveness.