Fairness-Based Global Optimization of User-Centric Networks

In this paper, we discuss the opportunities for global optimization of networks, which takes the user perspective into account. We state three fundamental problems in this approach, which are the variable (and usually also large) number of objectives, the Arrow impossibility theorem, and the unavoidance of the fact that the feasible space may contain solutions, where some objectives are extreme. To overcome these problems, we propose to achieve at least fairness among the fulfillment of user objectives, if optimality (e.g. in the sense of Pareto-optimality) is not possible. We also outline a generalized approach to fairness, based on fairness relation. Heuristic methods are available to handle the corresponding search problems.