A self-organized message ferrying algorithm

This work presents a self-organized message-ferrying algorithm called SOFCOM, which solves the next-hop decision problem in mobility-controlled message ferrying networks. We provide an analytic model for the single-ferry scenario, in which multiple static network nodes require transient connectivity for message exchange. Relaxing the system´s state complexity allows us to compare our approach with an optimal guessing oracle, with random-walk, and with a standard TSP-based approach. Based on the computed candidate solutions, we identify the type and the parameters of the solution space density function in order to judge our algorithm´s fitness. We show that SOFCOM can perform as good as a TSP-based approach, although only local decisions are conducted. We discuss important qualitative properties of our algorithm and provide a performance comparison.