Model Usage

We will exemplify the model assuming that it is used to represent the interaction between buyers and sellers in an online B2C marketplace. Marketplaces are the sellers, and individuals shopping at the website of the marketplaces are the buyers. Another example of a possible representation: hospitals or local markets like Carrefour vs. its users/clients.

A realistic consideration of buyers’ decision-making mechanism is done, based on the endorsement scheme and supported by data. Only buyers use endorsements. In a general consideration, the interaction is as follows. On the one hand, each time iteration, a set of agents (e.g., sellers) show a set of attributes that are observed and endorsed by another set of agents (e.g., buyers). This last set of agents learns (we can refer to them as the “set of agents that learn”), while interacting with the first set of agents, improving their decisions over time, by using endorsements. The first set of agents presents their attributes randomly. These are the attributes endorsed by buyers. The randomness of the attributes, is determined in accordance with how it is perceived in the real system by buyers. Diverse data sources can be used, e.g., surveys. At each iteration, buyers also interchange information via word of mouth (WOM). WOM allows buyers to communicate (suggest) the names of marketplaces to other buyers (each buyer suggests the best evaluated marketplace). On the other hand, at each iteration, buyers evaluate sellers (using the endorsement function to appraise the shown attributes) and choose a seller. Then, they endorse the chosen seller in accordance to the attributes it presents. Finally, sellers share information with others (via WOM), and the iteration finishes. The whole simulation process is illustrated in the figure shown above. Additional detail is given in the ODD summary table given below.