Abstract: Providing for record filtering in distributed dynamic clustering algorithms for coordinated multipoint (CoMP) wireless communication is described herein. By way of example, strategy selection records distributed as part of a belief propagation network are pruned at recipient nodes, thereby reducing processing overhead for dynamic clustering. As a result, cooperative policies can be determined with greater efficiency, and with greater relevance to local clusters of cooperating base stations. In some aspects, record pruning can comprise identifying and discarding redundant or incompatible sets of policy decisions. In at least one aspect, a number of evaluated records can be capped based on relevance, while preserving deployment-wide applicability of the belief propagation network. Accordingly, dynamic distributed CoMP decisions are optimized on a deployment-wide scale that more efficiently converges to maximum utility solutions.

Abstract: Systems and methodologies are described that facilitate cooperation strategy selection for a network multiple-in-multiple-out (N-MIMO) communication system. As described herein, one or more nodes in a communication system capable of N-MIMO communication can calculate marginal utilities, projected per-user rates, and/or other parameters corresponding to respective associated users. Based on these calculations, respective network nodes can perform user scheduling and selection, cell scheduling and selection, selection of a cooperation strategy (e.g., coordinated silencing, joint transmission, coordinated beamforming, etc.), and/or other operations to provide cooperative communication for respective users. As further described herein, projected rate calculation for a given user can be adjusted based on processing or channel implementation loss associated with the user, interference nulling capability of the user, or other factors.

Abstract: Systems and methodologies are described that facilitate dynamically forming clusters in a wireless communication environment. A set of non-overlapping clusters can be formed dynamically over time and in a distributed manner. Each of the clusters can include a set of base stations and a set of mobile devices. The clusters can be yielded based upon a set of local strategies selected by base stations across the network converged upon through message passing. For example, each base station can select a particular local strategy as a function of time based upon network-wide utility estimates respectively conditioned upon implementation of the particular local strategy and disparate possible local strategies that can cover the corresponding base station. Moreover, operation within each of the clusters can be coordinated.