Optimal Power Splitters' Locations and Cables' Routing in Distributed Antenna System

Due to the exponential growth in the demand of higher data rates in wireless systems, ensuring signal strength and efficient coverage has become a challenge facing any network operator. Distributed Antenna System is the solution deployed in this paper because when a sufficient number of distributed antennas is placed inside the building the coverage is improved, as radio signals can successfully reach any point in the building. Since improving the coverage increases the deployment cost due to the need of more antennas and equipment that will lead to more maintenance and installation cost, and minimizing the cost will actually lower the coverage, the aim of the paper is to formulate and solve the problem of finding the optimal cables' routing and the optimal power splitters' locations that connect all the antennas with the base station producing the minimum cost and satisfying the power requirements for each antenna at the same time. Problem solution delivers the tree topology of Distributed Antenna System components connection producing the minimum cost. The base station is considered the root of the tree while the antennas are the leaves. Power splitters are the nodes connecting the base station with the antennas. The objective function is minimizing the cost, it consists of linear terms as the cost of the cables and nonlinear terms as the installation cost of cables as there are two variables multiplied together. First, the variable that checks if the cable connects two components or not, and second, the variable that checks if the cable is installed before or not. The problem is simplified first to be solved without the nonlinear terms using integer linear programming tool and then a hybrid approach of integer linear programming and add heuristic algorithm is proposed to be solved with. Exhaustive algorithm was used as a reference for both algorithms. By comparing the integer linear algorithm and the hybrid approach, it was found that the hybrid approach delivers optimal solution in small buildings and near optimal solutions in large buildings. Second, after the nonlinear terms are included, the problem is solved at the beginning linear without considering the cable installation as a variable then the whole problem including the cable installation variable is solved using the proposed hybrid approach. Results show that the hybrid approach gives better solution than the linear solution after cable installation variable is considered in the formulation.