Simulated annealing user scheduling for coordinated heterogeneous MIMO networks

Abstract

In this work, a reduced complexity simulated annealing user scheduling algorithm to maximize the sum rate for the downlink of a coordinated multicell multiple-input multiple-output (MIMO) system is proposed. The users' data signals are transmitted cooperatively via the base stations of a cluster within the sectorized multicell heterogeneous cellular network. Assuming perfect channel state information, the performance of this system with the proposed algorithm is investigated. The simulations demonstrate the effectiveness of the proposed method in terms of sum rate, which is very close to the performance of the optimal but highly complex exhaustive search at much lower complexity.

Rotating clustering with simulated annealing user scheduling for coordinated heterogeneous MIMO cellular networks

“…We assume for ease of notation that and are both equal to . Similarly to [18], the definition for the aggregate downlink channel of the th user from all BSs in cluster denoted by is valid here when in [18] is replaced by here. The downlink channel matrix of the th user and th BS in cluster accounts for path loss, log-normal shadowing and Rayleigh fading.…”

“…The transmission for , is intra-cluster interference for user as given by the two summations in (3). Following a similar procedure for SZF-DPC as in [18], for a given ordered user , its achievable rate is given by (4) where is the transmit covariance matrix for user in cluster and are orthonormal basis vectors for the joint null space of the channels of all users encoded before in the encoding order. The effective channel matrix for user in cluster can be considered to be .…”

“…Similarly by defining and as, respectively, the data vector of user in cluster and its corresponding aggregate precoding matrix from all BSs (cf. [18]), the received signal for user is given by (1). The first term in (1) is the received signal from cluster , to which the user belongs, while the second term describes the interference from other clusters.…”

“…The sum of the ICI and noise in (1) is denoted by . Using a post-processing interference-whitening filter applied to the received signal (1) as described in [18], the equivalent post-processed signal can be written by again replacing by in [18, eqn. (4)]:…”

“…As a precoding method of interest, we use successive zero-forcing dirty paper coding (SZF-DPC), which partially nulls the interference between users [17]. Since reduced-complexity user scheduling is necessary in practical systems with a large number of users, we use a slightly modified version of the simulated annealing user scheduling (SAS) method that we used in [18], which achieves performance close to that of an exhaustive search with much lower complexity. We wish to emphasize, though, that the focus of this work is on the cluster rotation and its performance, not on the scheduling algorithm.…”