A Mixed-Integer Linear Programming framework for real-time actuator allocation in tokamak control

Actuator Management is a core high-level component of modern tokamak control systems, providing an interface between control algorithms and physical actuators. Its role includes distributing the aggregated controller commands among available actuators, substituting failed actuators, and dynamically assigning the most suitable actuators to control tasks in real time. The latter functionality is commonly formulated as an optimization problem, whose solution must be both computationally efficient for real-time application and capable of converging to the global optimum. This paper proposes a formulation of the actuator allocation problem within a Mixed-Integer Linear Programming framework and demonstrates how various factors discussed in the existing literature can be incorporated. It is shown that the resulting optimization problem can be solved within time limits compatible with real-time operation for large-scale systems such as ITER, even when executed on a standard personal computer, while providing guarantees of convergence to the global optimum.