This project is in collaboration with the companies Rio Tinto (RT) and Hydro-Québec (HQ). The former, a world leader in aluminium production, produces 90% of the energy required for its production via six hydroelectric power stations. The latter produces 98% of the province's electricity through its vast hydroelectric network. Both companies face many optimisation problems. Some of them cannot be addressed by existing operational research algorithms and methodologies. The main objective of this project is to develop new algorithmic approaches targeting some specificities identified in industrial optimisation problems. These approaches will be developed at the École Polytechnique de Montréal (ÉPM) by three students and two research associates, supervised by two professors who, in the past, have demonstrated their ability to train several graduate students, and have published numerous articles in the best optimization journals and in a wide range of engineering journals. The approaches developed in this project will then be integrated into the NOMAD software, an implementation of the MADS algorithm for black-box optimisation. The new algorithms can then be tested on the industrial partners' applications in collaboration with the researchers involved. The algorithmic progress will be published in scientific journals, and only the numerical results approved by the industrial partners RT and HQ will be released to the scientific community. The students involved will participate in all stages of the project and in particular in the exchanges with the industrial partners.