Swarming Algorithms for Distributed Radio Resource Allocation: A Further Step in the Direction of an Ever-Deeper Synergism Between Biological Mathematical Modeling and Signal Processing

Abstract

The major challenge faced by modern communication systems is the striking contrast between the ever-increasing demand for higher data rate links with guaranteed quality of service (QoS), on the one hand, and the scarcity of the radio resources, on the other hand. The solution of this dilemma is the increase of spectral efficiency. Several tools are available for improving spectral efficiency, like adaptive modulation and coding; multiple input, multiple output (MIMO) communications; cooperative multipoint communications; etc. Among all these tools, it is widely recognized that the approach having the potential for bringing the most significant improvement is spatial reuse of radio resources [1]. This calls for the deployment of base stations covering small cells of radius in the order of a few tens of meters. However, a dense deployment of conventional base stations is impractical because of the high costs of installation and maintenance. The way to overcome this limitation consists of the deployment of heterogeneous networks, which are composed of conventional base stations coexisting with low-cost base stations having very small transmit power and limited complexity. An example is given by femtocell networks [2], where femto access points, also called home enhanced node B, are installed indoors and cover cells of radius in the order of tens of meters. They avoid the wall penetration losses occurring in outdoor-to-indoor communications and are perfectly compliant with cellular standards. Consequently, they allow a significant traffic offload from the wireless to the wired channels, thus freeing important radio resources. This is a win-win strategy for operators and users: The operators get the capability of having more wireless traffic within a given area, thus increasing revenues; the users have a seamless connectivity to the network, with better QoS with respect to standard Wi-Fi and easier hand off. However, this solution does not come without technical challenges, the most important being interference management. In fact, while in conventional networks radio resource allocation is typically performed by the base stations, using a centralized approach, in the heterogeneous scenario, it becomes quite problematic to envisage a centralized approach taking into account both conventional and innovative base stations. The major limitation concerning the home base stations is that they have low complexity and, being owned by the users, they are not necessarily installed according to a radio optimization criterion and they may be switched on and off at the user?s will. Given this context, the only viable mechanism is to endow the network with some sort of self-organization capability, with a particular attention to keep interference under control. Self-organization can come in different forms: self-configuration, self-optimization, self-healing, etc. In particular, since some of the nodes have limited complexity and computational capabilities, it is fundamental to devise mechanisms such that the overall network is robust, as a whole, against abrupt and unpredictable changes of network conditions, like interference, change of connectivity, and so on, even if the individual nodes are not sophisticated enough to handle the problem appropriately. In fact, self-organization features are already introduced in the current standardization process of the 3rd Generation Partnership Project long-term evolution leading towards fourth-generation mobile communication systems [3].