Skip to main content

Table 1 Network dimensioning methodology (configuration #1)

From: Network dimensioning and base station on/off switching strategies for sustainable deployments in remote areas

  1. a This is an integer-programming problem for which analytic expressions of the blocking probabilities cannot be derived and, therefore, they have to be calculated numerically. In general, this kind of problems are NP complete, which means that optimality can only be achieved with a complexity similar or equal to exhaustive search. Only in some cases, when the analytic expressions of the constraint and objective functions are available (which is not the case in this paper), more simple methods can be derived with polynomial complexity but still assuring optimality. Note that the dimensioning problem we are facing has to be solved offline before the network is set up. A suboptimal solution implies that more resources than really needed will be deployed. There is no reason to stick to a suboptimal solution for the sake of computing such a solution faster, specially taking into account that in the remote regions that we are considering, the number of combinations to be checked by exhaustive search is relatively low and, therefore, the optimum solution can be found in just a few seconds or minutes according to the simulations carried out. Anyway, if the numerical complexity increases so much that suboptimal numerical methods have to be applied, this could be done by resorting to well-known techniques available in the literature. Some examples of algorithms fitting this problem (where the analytic expressions of the constraint and objective functions are not available) are genetic algorithms [35] and simulated annealing [36]