In onshore oilfields, sucker rod pumps are deployed to lift fluids from wells to the surface, which are then transferred by a pipeline network to a central facility. Each pump operates according to a cyclic control policy that alternates between a period of production and a period of recess, allowing the reservoir to replenish the fluids. To minimize the peak in the electric power drawn by the sucker rods, previous works have considered the scheduling of the control policies over time. However, when the production system sprawls over a large area, the energy required to force the flow through the pipelines and the pressure drops are significant. To this end, this article develops mixed-integer nonlinear formulations and mixed-integer linear approximations for baseline control-policy scheduling and reconfiguration.