The influence of seller learning and time constraints on sequential bargaining in an artificial perishable goods market

In this paper we test some hypotheses about the formation of prices through sequential bilateral bargaining on a perishable goods market under the assumption of behavioural learning by buyers and sellers. We constructed a multi-agent simulation model based on a survey concerning the fruit and vegetables market in Marseille, France. In our model, the agents bargain the price of a perishable good. The representation of agents' rationality is inspired by a literature relative to markets of perishable goods (Kirman [6] [7] and Rouchier [18] [19] [20]) as well as on artificial bargaining (Brenner[2], Weisbuch[22]). We study the influence of three parameters (the sellers' initial beliefs concerning the buyers' willingness to pay, the time spent by buyers on the market, and the heterogeneity in the sellers' limit value) on the evolution of prices and of agents' representations of others agents. We compare our results with empirical observations and some existing literature on bargaining. We find that our assumptions concerning the learning process give consistent results and lead the transactions prices to converge toward the sellers' limit value.