Competition among Payment Networks using Generalized Population Based Incremental Learning

We have developed an agent-based model for competition among payment cards networks. In our model the competitors in the payment card market learn by observing model-based interactions between costumers and merchants at the point of sale (POS). The interactions are represented on lattice with three different connections: local, small world and random. We are studying how the payment card providers improve their strategies in a competitive market. We are using Generalized Population Based Incremental Learning (GPBIL) algorithm as our machine learning technique to evolve strategies for card purveyors. In our Computational Agent-based model of Competition in the Payment Card Market, we are simulating the interactions among consumers and merchants in a way, which to our knowledge has not been explored previously in the literature. The simulation allows us explicitly represent the network externalities in the use/acceptance of payment cards and model its impact on the consumers/merchants decisions to adopt or drop a particular electronic payment. Additionally the decisions of consumers and merchants regarding the subscription and use of electronic cards are guided by the cost of the payment instruments, which is determinate by the payment card providers.