Application of extremal optimization algorithm to multi-objective topology design of enterprise networks

Extremal optimization (EO) is an optimization technique that has been utilized to efficiently solve many complex optimization problems. One such complex problem is topology design of enterprise networks which involves simultaneous optimization of a number of objectives. Some important objectives are financial cost, network delay, maximum number of hops between communicating nodes in the network, and network reliability. The problem also involves a number of design constraints. This paper proposes an EO algorithm to efficiently solve the topology design problem of enterprise network. The multi-objective aspects of the problem are managed through the incorporation of goal programming method in the EO. A preliminary analysis of the proposed EO is presented, and a comparison is done with simulated annealing algorithm which is a well-established optimization algorithm. A program was developed in C++ for simulating the test conditions. Empirical results show the suitability of EO for the problem. Furthermore, EO produced solutions of better quality when compared with the simulated annealing algorithm.