Heuristic algorithms in topological design of telecommunication networks

Abstract

The paper addresses the generic topological network design problem and considers the use of various heuristic algorithms for solving the problem. The target of the optimisation is to determine a network structure and demand allocation pattern that would minimise the cost of the network, which is given by fixed installation costs of nodes and links and variable link capacity costs described by linear or concave functions. Input data for the optimisation consists of a list of potential node and link locations and their costs and a set of demands defined between the nodes. Since the problem is known to be NP-hard, the use of specialised heuristic algorithms is proposed. The presented approaches encompass original ideas as well as selected methods described in literature and their enhancements. The algorithms are based on the following ideas and methods: shifting of individual flows, local and global restoration of flows from chosen links or nodes, Yaged algorithm for finding local minima, Minoux greedy algorithm, simulated allocation and genetic algorithms. Efficiency of each of the proposed methods is tested on a set of numerical examples.