Minimalist Self-Organization in Wireless Networks

Abstract

Many fields of human endeavour, such as biology and the theory of complex systems, are now embracing theconcept of self-organization based on local actions leading to a desirable global emergent behavior. While many examples,both natural and artificial, can be found of such self-organized systems, the relationship between the local rules and the globalbehavior remains elusive and no systematic procedure is known to engineer a specific global result. Given the increasingpervasiveness of wireless networks of all sorts, including ad hoc networks competing within narrow unlicensed bandsand wireless sensor networks, self-organization could constitute the next defining paradigm in wireless communications. Itcan be shown that a set of heuristic principles can be leveraged to engineer a self-organized connection-oriented wirelessnetwork with minimal complexity. Such a system requires no centralization of information, yet achieves a nearly optimalglobal state with only a modest amount of local signaling. It will naturally and jointly balance the many parameters relatedto radio resource management, exhibiting great adaptability, fault tolerance and scalability.