Optimal New Site Deployment Algorithm for Heterogeneous Cellular Networks

With exponential growth of mobile broadband traffic, mobile operators are facing the challenge of cost-efficiently evolving 3G/4G network to meet future network capacity demand. Heterogeneous multilayer network deployment is one essential performance-effective and potentially cost-effective candidate solution. We study optimal new site deployment algorithms in heterogeneous wireless cellular networks. We focus on a joint Macro and Micro layer new site deployment. We propose and evaluate a class of fast-convergent meta-heuristic algorithms: Scalable Multi-radio deployment AlgoRiThm (SMART). Their performances and convergence speeds are evaluated using a large-scale real dense-urban HSPA network deployment. We compare the performance on four variants of SMART: Greedy, Simulated Annealing, Greedy with memory and Simulated Annealing with memory. We show that the Simulated Annealing and Greedy algorithms achieve almost the same optimal deployment when the number of optimization iterations is large and, the Greedy algorithm converges much faster than Simulated Annealing algorithm when the number of optimization iterations is small. We also show that in a practical network deployment scenario SMART outperforms previous heuristic deployment algorithms proposed in the literature.