Hierarchic Topology Management by Decision Model and Smart Agents in Space Information Networks

Space information network, which is envisioned as a new type of self-organizing networks constituted by information systems of land, sea, air, and space, has attracted tremendous interest recently. In this paper, to improve the data delivery performance and the network scalability of space information networks, a new hierarchic topology management scheme based on decision model and smart agents is proposed. Different from the schemes studied in mobile ad hoc networks and wireless sensor networks, the proposed algorithm in space information networks introduces a decision model based on analytic hierarchy process (AHP) to first select cluster heads, and then forms non-overlapping k-hop clusters. The proposed dynamical self-maintenance mechanisms take not only the node mobility but also the cluster equalization into consideration. Smart mobile agents are used to migrate and duplicate functions of cluster heads in a recruiting way, besides of cluster merger/partition disposal, reaffiliation management and adaptive adjustment of information update period. Simulation experiments are performed to evaluate the performance of the proposed algorithm in terms of network scalability, overhead of clustering and reaffiliation frequency. It is shown from the analytical and simulation results that the proposed hierarchic topology management algorithm significantly improves the performance and the scalability of space information networks.