A multiobjective approach to resource management in smart grid

A multiobjective (MO) approach is proposed for resource management in the smart grid. The resource management problem is formulated via a state-space model. The grid performance is optimized and the distribution cost of resources is minimized, yielding an MO optimization problem. An H
_∞
design is adopted to address grid disturbances, and linear matrix inequality approaches are used to render the MO problem numerically solvable. An algorithm that integrates deterministic and stochastic mechanisms is presented. By using the algorithm, an approximated Pareto front (APF) can be obtained. The resources in the smart grid can thus be managed by choosing one of the trade-off strategies represented by nondominated vectors on the APF. In contrast with a single-objective approach, the proposed MO approach provides a grid designer with a broad perspective on optimality, which clearly illustrates how one objective affects the other. Simulations show that by using the MO approach, the distribution cost can be significantly reduced with moderate degradation of the grid performance.