A decentralized control strategy for energy retrofit planning of large-scale street lighting systems using dynamic programming

This paper addresses the energy management of large-scale urban street lighting systems. We propose a multi-stage decision-making procedure that supports the energy manager in determining the optimal energy retrofit plan of an existing public street lighting system. The problem statement is based on a quadratic integer programming formulation and aims at simultaneously reducing the energy consumption, ensuring an optimal allocation of the retrofit actions, and efficiently using the available budget. The proposed solution relies on a decentralized optimization algorithm that is based on discrete dynamic programming. The methodology is applied to a real street lighting system in the city of Bari, Italy.