Topology formulation algorithms for wireless networks with reconfigurable directional links

In future airborne networks, high rate connectivity between platforms will be heterogeneous and provided over directional free space optical (FSO) or RF wireless links. It is essential for large scale high performance networks that the creation and maintenance of link adjacencies between airborne nodes proceed in an automated manner. Precisely which steerable transmit beams should form connections with available receivers within the transmitter´s field of view, range, and type must be determined-quickly. To accomplish this aim we seek an algorithm to create a robust network that balances stability and performance goals given connectivity and resource constraints. Due to the unstable nature of links and data traffic patterns, the primary objective is to maximize the number of connected nodes, an NP-hard problem. In this paper, we describe an integer linear programming (ILP) formulation that discovers a degree constrained spanning tree (dST) if one exists. If no dST exists, i.e., if no degree constrained connected graph is feasible, the algorithm produces a degree constrained forest with the maximum number of edges feasible, subject to the requirement that each tree contains at least one edge. Performance and scalability results are shown for the ILP variant described, in different scenarios and for different network sizes.