Optimization approaches for planning small cell locations in load-coupled heterogeneous LTE networks

A key planning task in heterogeneous networks is to optimally plan small cell locations for capacity enhancement and offloading. In LTE systems, the cells' load levels are inherently coupled with each other in a non-linear manner because of mutual interference, which complicates the task. In this paper, we address the offloading aspect by optimally selecting, from a set of candidate locations, up to a given number of small cell sites in a deployed macro cell network, subject to the load-coupling relation between the cells. We prove that the problem is NP-hard and present computational optimization approaches. The first approach is a search algorithm targeting near-optimal solutions of deploying small cells. The second, complementary, approach aims at numerically constructing a tight linear approximation of load coupling, with a possibility of iterative refinement, to enable the use of mixed-integer linear programming to gauge optimality. The optimization concepts are illustrated using a representative LTE heterogeneous network scenario. The results demonstrate the effectiveness of the optimization approaches, and enable insights into optimal offloading with respect to traffic demand and the number of small cells.