Strategic Open Routing in Service Systems with a Shared Station

An accurate prediction of the behaviour of strategic individuals in an open routing service system with a shared station can be of great use for many companies, an example is a restaurant with an open buffet. However, customer behaviour is hard to predict and system details tend to differ a lot. In the paper of Arlotto et al. (2019) an open routing system with two stations is simulated. They conclude customers show herding behaviour at the slowest station, to prevent being further back in the slow queue, as this could happen if they visit the faster station first. In this research, we first replicate their simulation and compare the results. In addition, we perform a similar simulation on a three-station subset open routing system with a shared station. Customers in this station visit only two of the three stations, but the shared station is always one of them. We divide the customers into two groups, one group visits one subset of the stations and the other group visits the other subset. The simulation shows individuals tend to herd at the shared station, given high enough service rates at the two specific stations.