Location models to optimize flood monitoring networks

• Autor(a) principal: Simoyama, Felipe
• Data de Publicação: 2023
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Repositório Institucional da UNIFESP
• Texto Completo: https://repositorio.unifesp.br/11600/69543
• topic: Optimization; Location models; Flood monitoring; Rain gauge; Maximum coverage

Floods are the most common type of disaster worldwide, accounting for approximately 43% of occurrences of disasters over the past 20 years. Besides the loss of hundreds of thousands of lives globally, it is estimated that over 600 billion dollars in damages have been caused by floods, including damage to infrastructure such as homes, schools, and hospitals. Concurrently, floods have significant impacts on urban mobility. For example, when a road is flooded, it not only affects local residents but also those who use that road and its surroundings for transportation. Furthermore, due to logistical issues and their economic consequences, these problems indirectly affect even those who do not pass through the flooded area. Flash floods are often triggered by extreme rainfall events that occur in areas with drainage problems, such as urban roadways. To minimize the impacts of flooding, transportation departments in various cities and metropolitan regions around the world use sensor data to issue alerts and take actions to mitigate these impacts. One of the most commonly used devices is the rain gauge, as it provides point-specific and factual information about rainfall. These sensors are generally not deployed in isolation but rather forming a network. The placement of these networked sensors is a widely discussed problem in the literature, especially in the last decade, with the number of publications increasing year by year. The four studies that constitute this thesis specifically address this problem through the lenses of operations research. A systematic literature review was conducted on the optimization of rain gauge networks, along with three applications of location models for this problem. These studies, presented in their entirety in this thesis, demonstrate the importance of adjusting classical location models to the specific problem, given its intrinsic uncertainties and the uniqueness of each study area.