The transition to a sustainable energy future requires innovative solutions that reduce greenhouse gas emissions across various sectors. In this context, hydrogen has emerged as a promising alternative fuel, particularly for applications where electrification is challenging. One such sector is fisheries, which heavily relies on diesel-powered vessels and equipment.

Green hydrogen is produced by electrolysis and transported to end users by truck or ship. Currently, hydrogen is transported in compressed form with low energy density. Therefore, it is beneficial to reduce transport costs by implementing distributed production. However, this cost-saving approach needs to be balanced against the more efficient production and lower costs of larger electrolysers. The electrolysers are assumed grid-connected and may be exposed to time-varying electricity and power tariffs. Additionally, they produce surplus heat that can be sold for commercial value.

To integrate all these aspects into one model, we use the JuMP modelling framework in Julia. Our modular approach allows us to set up and solve a mixed-integer linear optimization problem for determining the location and dimensions of the hydrogen infrastructure. To address non-linearities arising from varying electrolyser efficiency and dynamics in the electricity market, we employ a package that identifies optimal piecewise linear convex approximations based on a set of data points. This approximation of production costs is then combined with separate modules for storage and transport modeling, resulting in an overall facility location model.

Initial results from a case study in the ZeroKyst project based on a region of Northern Norway will be used to illustrate the approach.

Given an allotment of land divided into parcels, government decision-makers, private developers, and conservation biologists can collaborate to select which parcels to protect, in order to accomplish sustainable ecological goals with various constraints. In this talk, we propose several mixed-integer optimization models that considers the presence of multiple species on these parcels, subject to predator-prey relationships and crowding effects. Recognizing that the relationships within and across species, their resources, and environment are quite complex, we also incorporate simulation-based estimates of long-term population viability into our models.