Cooperative decision-making under uncertainty: applications to energy markets and railway systems
This dissertation develops mathematical models for decision-making under uncertainty, withapplications in renewable energy markets, cooperative decision-making, and railway systems. Itinvestigates how effective decisions can be designed when future outcomes are uncertain,information is incomplete, or system parameters cannot be known precisely.
For wind power producers, uncertain electricity generation creates a direct financial challengebecause market commitments must be made before actual production is known. This dissertationdevelops stochastic models and multi-period bidding strategies to examine how producers canbalance expected profit against exposure to imbalance risk. The results show that incorporatingstorage information and accounting for decisions across multiple market periods can improveeconomic performance while reducing financial risk.
The research then develops cooperative models for situations in which multiple decision-makersform coalitions and coordinate their decisions under uncertainty. These models quantify howcooperation and aggregation can reduce variability and generate additional economic value.Allocation mechanisms are designed to distribute these benefits in a fair and stable manner.Moreover, learning-based models are developed that allow joint decisions and profit allocationsto adapt progressively as new information becomes available.
Finally, this dissertation develops control models for virtually coupled rail vehicles, in whichvehicles coordinate their speeds and relative distances through communication and feedback.The analysis accounts for uncertainty in physical factors such as train mass, track slopes, andaerodynamic forces, and establishes conditions under which coordinated operation remainsstable and reliable.
Overall, this dissertation advances mathematical methods for decision-making under uncertaintythrough optimisation, cooperative game theory, learning, and control.