Data-driven optimization and statistical modeling to improve meter reading for utility companies

Utility companies collect usage data from meters on a regular basis. The are collected automatically using radio-frequency identification (RFID) technology. Each meter transmits signals an RFID tag that read by vehicle-mounted reading device within specified distance. Routing the vehicles can be modeled close-enough vehicle routing problem street network. In practice, there is uncertainty while meters. signal transmitted discontinuous, and range each different stochastic due to weather conditions, surrounding obstacles, interference, decreasing battery life of tags. These factors lead not being read. A has sent at later time missed meters, this leads increased costs for utility company additional operational overtime payments drivers. Our aim address issues technology generating routes both cost-effective robust (we seek minimize number reads). We use analytics, optimization, Bayesian statistical models uncertainty. Simulation experiments real show hierarchical model gives better results compared other models. potentially integrate into their route software as decision-support tool produce more than they currently generate. • Formulate two-stage integer program (IP). IP formulation deterministic even though stochastic. Develop three learning capture inherent in data. performs non-hierarchical iterative algorithmic framework directly use.