Article ID: | iaor20061471 |
Country: | Netherlands |
Volume: | 164 |
Issue: | 2 |
Start Page Number: | 301 |
End Page Number: | 323 |
Publication Date: | Jul 2005 |
Journal: | European Journal of Operational Research |
Authors: | Batta Rajan, Rogerson Peter A., Akella Mohan R., Delmelle Eric M., Blatt Alan, Wilson Glenn |
Keywords: | location, health services |
The location of base stations (BS) and the allocation of channels are of paramount importance for the performance of cellular radio networks. Also cellular service providers are now being driven by the goal to enhance performance, particularly as it relates to the receipt and transmission of emergency crash notification messages generated by automobile telematics systems. In this paper, a Mixed Integer Programming (MIP) problem is proposed, which integrates into the same model the base station location problem, the frequency channel assignment problem and the emergency notification problem. The purpose of unifying these three problems in the same model is to treat the tradeoffs among them, providing a higher quality solution to the cellular system design. Some properties of the formulation are proposed that give us more insight into the problem structure. An instance generator is developed that randomly creates test problems. A few greedy heuristics are proposed to obtain quick solutions that turn out to be very good in some cases. To further improve the optimality gap, we develop a Lagrangean heuristic technique that builds on the solution obtained by the greedy heuristics. Finally, the performance of these methods is analyzed by extensive numerical tests and a sample case study is presented.