Doctoral Position in Robust Railway Intervention Planning under Uncertainty — ETH Zürich

Full description

Doctoral Position in Robust Railway Intervention Planning under Uncertainty

100%, Zurich, fixed-term

print Drucken

The Chair of Infrastructure Management, led by Professor Dr. Bryan T. Adey within the Institute of Construction and Infrastructure Management of the Department of Civil, Environmental and Geomatic Engineering, has an opening for a doctoral student. This position focuses on the development of uncertainty-aware methods and optimisation tools for robust railway intervention planning and resource forecasting. The position is connected to the interdisciplinary ETH Mobility Initiative research project conducted in collaboration with the SBB.

Project background

Railway infrastructure provides substantial capacity for the movement of people and goods, yet this capacity is reduced or suspended whenever there are interventions. Interventions, which range from condition monitoring through minor and major maintenance to renewal and expansion, require time, money, machines, and personnel, and they partially block access to the track. Consequently, traffic and timetables must be modified, and passengers are affected. Decisions on when, where, and how such interventions should be grouped or separated, and on how the associated timetables should be planned, must be taken well ahead of execution and under considerable uncertainty regarding maintenance costs, timetable feasibility, and passenger impact.

These decisions are challenging because the relevant effects and factors are difficult to quantify and are inherently uncertain, whilst decision-making power is distributed across several stakeholders, asset levels, and time horizons. The coordination process is at present largely qualitative and iterative, and it offers limited scope for the systematic use of predictive, quantified information. Multiple trade-offs must therefore be balanced, including direct economic costs, the availability of contractor resources, short-term effects on passengers such as longer journeys and additional transfers, and long-term effects such as the erosion of trust and of political support for railway funding.

This ETH Mobility Initiative project addresses these gaps by developing quantitative support for fact-based decision-making in railway infrastructure management. Particular emphasis is placed on characterising and propagating the uncertainties inherent in intervention planning and resource forecasting, on developing robust optimisation methods that determine when, where, and how interventions should be grouped or separated, and on integrating the resulting information within a geospatial decision-support environment aligned with ISO 55001 and UIC best practices. The advertised position contributes to this line of work, with validation carried out against historical and planned data for a pilot SBB corridor.

Job description

This doctorate aims to develop uncertainty-aware methods and optimisation algorithms for robust intervention planning and resource forecasting in railway infrastructure management.

• Characterising and propagating the principal sources of uncertainty in intervention planning and resource usage across multiple time horizons and asset categories
• Extending mixed-integer linear programming models through stochastic and robust optimisation to compute robust intervention programs
• Contributing to the design of a GIS-based decision-support platform that integrates asset data, intervention plans, and uncertainty analytics
• Validating the developed methods against historical and planned SBB corridor data

Profile

• A Master’s degree in civil engineering, transport engineering, operations research, applied mathematics, computer science, or a closely related discipline
• A strong background in mathematical optimisation, in particular mixed-integer linear programming, together with stochastic or robust optimisation and solid programming skills (preferably in Python)
• Familiarity with uncertainty quantification,