Optimisation and Simulation for Decision Support
[Open soon]
Work description
The fellowship is part of the DFence – Dynamic Virtual Fences project, which aims at the autonomous definition of virtual fence boundaries for extensive grazing, based on computer vision, forecasting models and planning. The work aims to investigate, develop and validate the pasture-use planning, simulation and optimisation module of the project's Integrated Decision Support System. This module should generate actionable grazing plans that i) prevent over- and under-utilisation of pastures, ii) ensure that animals graze in the most suitable areas and iii) ensure sustainable management practices, serving as the basis for the dynamic definition of virtual fences. Optimisation models based on exact methods (e.g. mixed-integer linear programming) and approximate methods (e.g. metaheuristics, evolutionary algorithms, multi-objective optimisation) will be studied and developed, as well as agent-based and discrete-event simulation models. Simulation will make it possible to analyse the uncertainty (weather, external factors, animal behaviour) and variability (biomass availability and regrowth, per-animal intake) of the system, and to assess the performance of plans and forecasts in a stochastic environment as close to reality as possible. The module will take as inputs historical sensor and telemetry data, terrain and vegetation attributes extracted from imagery, and pasture-state forecasts produced in other project tasks. The work will build on existing project results: an agent-based simulation model of the study area and a data pipeline organised in layers (Bronze, Silver and Gold), with data products and data contracts, orchestrated workflows and quality reports. In addition to developing the planning module, the fellowship holder will be responsible for integrating it with the pipeline and for the continued maintenance and evolution of the pipeline, closing the data–forecast–simulation–optimisation–decision loop. Specifically, the main activities to be carried out by the fellowship holder are: • Study the state of the art in grazing planning and optimisation, agent-based and discrete-event simulation applied to livestock systems, optimisation under uncertainty, simulation-optimisation approaches, and virtual fencing. • Formalise the pasture-use planning problem (decision variables; spatial, water, topographic and animal-welfare constraints; productivity, ecological sustainability and welfare objectives), in coordination with the consortium partners. • Develop and compare exact and approximate optimisation models, including multi-objective formulations, to generate grazing plans and virtual fence configurations. • Extend the agent-based and discrete-event simulation models, incorporating uncertainty and variability, and use them to assess the robustness of the plans and the accuracy of the forecasting and optimisation methods in a stochastic environment. • Integrate the planning module with the data pipeline, automating the generation of scenarios and simulation configurations from the data products and the consumption of forecasts and image-derived attributes. • Maintain and evolve the data pipeline: incorporate new data sources and versions (e.g. GPS collar telemetry), update workflows, data products and data contracts in an additive and versioned way, monitor data quality and improve processing performance. • Calibrate and validate the models with real telemetry and field data, as they become available, and with the expert knowledge of the project partners. • Systematise the results of the work, producing technical documentation, reports and, where applicable, scientific contributions in the form of papers, datasets or reusable software. The training component will include the integration of the fellowship holder into a multidisciplinary research team, training in optimisation, simulation and data engineering tools, and participation in consortium technical meetings and science dissemination activities.
Academic Qualifications
Bachelor’s degree in Computer Engineering, Industrial Engineering and Management, Information Systems or a related field.
Minimum profile required
Bachelor’s degree final grade above 13.
Preference factors
- Proficiency in Portuguese and English (written and spoken). - Experience in developing and maintaining data pipelines (e.g. Python, SQL, workflow orchestration, analytical data formats). - Experience in optimisation modelling (mathematical programming and/or metaheuristics) and/or in agent-based or discrete-event simulation. - Experience in processing geospatial data (vector and raster). - Enrolment in a master's programme in a relevant field.
Application Period
Since 22 Oct 2026 to 04 Nov 2026
[Open soon]
Centre
Industrial & Systems Engineering and Management