This paper describes an application of multi-agentsystems to economics. The issue we address in ourwork is the actual functionning of a distributedmarket in terms of information acquisition for agents.We show here possible links between individualinformation acquisition and global results in thefluidity of exchanges. A multi-agent model has beenbuilt on the basis of field observations and interiewsled at the Marseille Fruits and Vegetables wholesalemarket in 2002-2003. Two types of artificial agentsinteract, wholesale sellers and retailers, surrounded byan environment of exogenous supply and prices.Retailers can choose wholesale sellers according totwo logic: either following some relational loyalty orsearching for best prices (which takes time). We studythe influence of the number of agents with eachrationality so that to see the importance of bothcategories in the good supply of consumers and thelimitation of garbage.

Keywords: Agent-Based Model, Multi-Agent Simulation, Market, Learning, heterogeneity, network.