Comparison of Scenario-Based and Interval Optimization Approaches to Stochastic SCUC

Abstract

This paper compares applications of scenario-based and interval optimization approaches to stochastic security-constrained unit commitment (Stochastic SCUC). The uncertainty of wind power generation is considered in this study to compare the two approaches, while other types of uncertainty can be addressed similarly. For the simulation of uncertainty, the scenario-based approach considers the Monte Carlo (MC) method, while lower and upper bounds are adopted in the interval optimization. The Stochastic SCUC problem is formulated as a mixed-integer linear programming (MIP) problem and solved using the two approaches. The scenario-based solutions are insensitive to the number of scenarios, but present additional computation burdens. The interval optimization solution requires less computation and automatically generates lower and upper bounds for the operation cost and generation dispatch, but its optimal solution is very sensitive to the uncertainty interval. The numerical results on a six-bus system and the modified IEEE 118-bus system show the attributes of the two approaches for solving the Stochastic SCUC problem. Several convergence acceleration options are also discussed for overcoming the computation obstacles in the scenario-based approach.