Modeling Competition among Wireless Service Providers

We consider a scenario where a population of customers is spatially distributed in a region which is served by two wireless service providers that offer Internet Access via two noninterfering technologies: one having a uniform coverage over the region (e.g. WAN), and the other, a limited coverage (e.g. WiFi “hotspots”). We assume that customers are equipped with “dual mode” wireless communication devices that have the capability to select which among the available providers to use. We introduce a stochastic geometric model for the locations of customers and providers’ access points and a utility-based mechanism modeling how devices select among providers. In particular, we assume that each device makes greedy decisions at random times, i.e., selects the available provider offering the highest utility at that time. We demonstrate that this process may have multiple equilibria, and prove that the system will almost surely evolve to one of the equilibrium configurations, starting from any initial configuration for users’ choices. We also provide conditions under which the set of equilibria is relatively “tight” – in this case the equilibrium configuration of agents’ choices is "maxmin fair" and thus is desirable if providers wish to cooperate in providing users with worst case performance guarantees. As an application of our framework we analyze the WAN and WiFi competition in an asymptotic scenario where the service zones of WAN provider are much larger than those of WiFi access providers.