Security constrained unit commitment（SCUC） and security constrained economical dispatch are the core component of generation schedule optimization， while robustness is the critical issue for its application in grid operation. Based on an analysis of a general security constrained unit commitment optimization model， a relaxed constraints model for security constrained generation schedule optimization is proposed. Through the relaxation of energy constraints， environmental constraints， network constraints and generation capacity constraints， the convergence performance of the algorithm is rapidly improved. A day-ahead and real-time generation schedule optimization application based on the model and algorithm is presented， which proved a success in a provincial power grid.