Energy consumption scheduling to achieve a low power generation cost and a low peak-to-average ratio load is a critical component in the next generation of power systems, known as smart grids. Implementing such a component requires the knowledge of the whole power demand throughout the system. However, due to the diversity of power demands, this requirement is not always satisfied in practical scenarios. To address this inconsistency, the present paper addresses energy consumption scheduling in a multi-grid power system (consisting of a local grid with some neighbor inaccessible grids). The total cost minimization is formulated as an optimization problem. In addition to optimal solution, the performed formulation is provided with online stochastic iterations to capture the randomness of unknown load adaptively over time. Numerical results demonstrate the effectiveness of the proposed algorithm in following the results obtained from the optimal solution.
