Optimization Problems Under Uncertainty in Smart Cities

Abstract

Combinatorial optimization has proven to be a valuable tool in several real case applications. With the advent of smart cities, it has naturally been leveraged by both researchers and industries to deal with problems such as optimal allocation of urban services, effective management of scarce resources, planning of logistic operations over congested networks, and pollution reduction. In practice, these problems are turned into optimization models, and their solution reveals the best tactical or operational policy able to maximize a particular utility or minimize a particular cost. Typically, several parameters involved in these models (e.g., people's behavior and market fluctuations) are influenced by many uncertain variables that are only partially or not known by the decision-maker. This lack of knowledge can be dealt with by using specific optimization paradigms such as the Robust Optimization or the Stochastic Programming. Nevertheless, the complexity of these models precludes, in most cases, to find an optimal solution. Thus, approximation methods have to be considered. This chapter presents several applications leveraging optimization models and algorithms for effective management of urban areas and their operations and a novel and efficient Extreme Values Theory-based deterministic approximation framework to cope with uncertainty in smart cities' optimization and planning.