Date of Award


Degree Type


Degree Name

Master of Science (MS)


Civil and Environmental Engineering

Committee Chair(s)

Gilberto Urroz


Gilberto Urroz


David Rosenberg


Sarah Null


How should a decision-maker choose which hydropower plant to build in a given river? Hydropower plant expansion is a task that involves several stakeholders, objectives, and resources. It requires that decision-makers use not only experience and previous knowledge to come up with their final decision, but also the use of tools to help them to decide, as well as to support their choices. The present project developed a computation model to optimize the choice of hydropower plants that can be built within a given river, considering three different objectives - energy generation, flooded surface area and river connectivity. The input data required to solve the mathematical problem was previously obtained from other sources. The mathematical model developed is classified as multi-objective linear integer problem and was solved using a the branch-and-cut method implemented by a commercial linear solver. In order to deal with the multi-objective approach, the constraint method was used. The main outputs of the model are the trade-off curves between objectives and the Pareto sets, which can be used by decision-makers to better understand the existing tradeoffs imposed by the problem. The model was tested using the energy generated the flooded surface area and distance from mouth of each project studied in the hydropower inventory studies developed for the Paru River, a tributary of the left margin of the Amazon River, located in the state of Pará, Brazil. The results showed that there is a trade-off between energy generation and surface flooded area, as well as energy generation and river connectivity, both a direct consequence of the fact that more projects will be necessary to meet energy requirements. Connectivity and surface flooded area, however, are not competing, since an improvement in both metrics will result in habitat quality improvement. The results were also compared to the ones obtained in the inventory studies, showing that different methods (ranking methods x multi-objective optimization models) can yield different policies. The main reasons for these differences are related to the solution method used since each one will take into account different management goals and performance metrics.