Anotace:
Today, passive optical networks (PONs) represent a modern solution for high-speed FTTx subscriber lines and networks. Due to the fact that optical fibres and cables are deployed in last-mile network segments. Unfortunately, the costs of trenching and installation of optical cables in cities are still very high today. However, the application of mathematical methods and algorithms for designing optimum network solutions and topologies might significantly decrease the overall capital expenditures (CAPEX) of the whole process. This article introduces an efficient method based on agglomerative clustering for an optimization of deployment of PONs in practice. The paper describes its application and proposes the necessary modifications of the clustering method for the environment of PONs. The proposed algorithm uses the technique to cluster and connect the adjacent optical terminals and splitters of PONs and to calculate the optimum locations of these units in order to achieve the most optimum network solution and to minimize the summary capital expenditures.