Interior-Point Methods for Optimal Power Flow with Scheduled Topological Changes under Uncertain Demand

Abstract Effective operation of modern power systems relies on optimization models that can account for both demand uncertainty and scheduled network reconfigurations. This paper presents a two-stage stochastic formulation of the pre-dispatch problem with scheduled line and bus switching, formulated within the linearized DC power flow framework. In the first stage, generation setpoints are scheduled before demand is realized; in the second stage, these decisions are adapted to multiple demand scenarios and the resulting topologies after switching. Network maneuvers are represented through a unified binary configuration matrix, which compactly represents simultaneous line and bus events while preserving the sparsity of the model. The resulting block-structured problem is solved efficiently with a specialized primal-dual interior-point method. The proposed approach integrates stochastic modeling, explicit representation of switching actions, and advanced solution techniques into a unified and scalable framework. By contrasting the method with chance-constrained DC-OPF, the paper highlights the advantages of scenario-based redispatch in capturing both uncertainty and topological flexibility. This work advances the state of the art in stochastic optimization for power system operation, providing a rigorous yet practical foundation for secure and economical decision making in increasingly uncertain and dynamic environments.