Uncertainty
Systemic, environmental and behavioural disruption, quantified as it grows across timescales.
Combining prediction with optimisation under uncertainty, from operations to long-term planning.

Systemic, environmental and behavioural disruption, quantified as it grows across timescales.
Traveller preferences and their adaptation to new mobility solutions, estimated and updated without re-running surveys.
Systems that absorb disruption and use scarce resources efficiently, remaining usable under varying conditions.
The closed-loop integration of estimate-then-optimise, in which operations and strategy inform one another.
Operational decisions and long-term plans are often made separately: real-time information about disruption rarely informs investment, and strategic plans do not adapt to daily conditions. TRANSFORM addresses this gap with a single methodological framework.
The framework combines machine learning, operations research and behavioural science in an estimate-then-optimise approach: predictive models quantify uncertainty as it evolves, and optimisation acts on those predictions across days, years and decades. Multimodal urban mobility is the setting in which the methods are developed and tested; the framework also applies to other systems that decide under uncertainty, including energy and healthcare.
Four themes, each pairing a methodological innovation with its expected outcome.
Disruption-aware supply forecasts that predict resource availability from noise, with probabilistic outputs that integrate directly into optimisation.
Reliable forecasting of resource availability across times and locations in a dynamically evolving network.
Behavioural adaptation to environmental change: choice parameters updated as new mobility solutions enter the network, without re-running surveys.
Behaviourally adaptive demand management through a user-centric multimodal trip recommender.
Adaptive multi-timescale optimisation that propagates uncertainty envelopes from operations through to strategic investment.
Dynamic, adaptive optimisation that links daily operations with long-term planning.
A move from agent-based to closed-loop multi-agent simulation, with temporal aspects and network adaptation embedded in the platform.
Multi-timescale, multi-agent simulation for assessing the impact of integrated decisions.