On the potential of using the cable trench problem in planning of ICT access networks

In this paper we are analyzing the potential winnings of solving a very hard combinatorial optimization problem in connection with dimensioning cables and trenches for deployment of green field information and communication technology access networks. Firstly we motivate the problem by summarizing the two traditional approaches used to solve the mentioned task, namely; the shortest path spanning tree and the Steiner minimal tree, which minimize cables and trenches respectively. Afterwards the cable trench problem is introduced followed by novel results on the potential winnings and possible reductions in relation to the problem. Finally it is concluded that a potential for using the cable trench problem exists, with cost savings in the order of 8% compared to the most efficient traditional approach. With respect to reductions it is concluded that these are not efficient enough to make the problem solvable using simple exact methods such as LP-relaxation and branch-and-bound.