Fast Lagrangian relaxation for constrained generation scheduling in a centralized electricity market

Abstract

This paper proposes a fast Lagrangian relaxation (FLR) for constrained generation scheduling (CGS) problem in a centralized electricity market. FLR minimizes the consumer payment rather than the total supply cost subject to the power balance, spinning reserve, transmission line, and generator operating constraints. FLR algorithm is improved by new initialization of Lagrangian multipliers and adaptive adjustment of Lagrangian multipliers. The adaptive subgradient method using high quality initial feasible multipliers requires much less number of iterations to converge, leading to a faster computational time. If congestion exists, the alleviating congestion index is proposed for congestion management. Finally, the unit decommitment is performed to prevent excessive spinning reserve. The FLR for CGS is tested on the 4 unit and the IEEE 24 bus reliability test systems. The proposed uniform electricity price results in a lower consumer payment than system marginal price based on uniformly fixed cost amortized allocation, non-uniform price, and electricity price incorporating side payment, leading to a lower electricity price. In addition, observations on objective functions, pricing scheme comparison and interpretation of Lagrangian multipliers are provided.