—Wireless mesh networks are becoming prevalent in providing access infrastructure to the public. Design of such networks involve several important goals including resilience to failures, cost minimization, and accessibility to a large number of users. In this paper, we propose Greedy Selection Rounding (GSR), an efficient and near-optimal algorithm to design a network topology that maximizes the coverage of the users while ensuring that the network is resilient to node failures, the deployed routers are load balanced, and the deployment cost is under a given budget. Furthermore, we extend GSR to address higher robustness requirements, heterogeneous deployment costs/profits, while bounding the path lengths. The GSR algorithm has the useful property that when it fails to find a solution satisfying the budget constraint, the incurred cost is at most a constant factor more than the desired budget. Through extensive evaluation, we show that GSR performs much better than its theoretical...
Chun-cheng Chen, Chandra Chekuri, D. Klabjan