A novel binary integer linear programming traceable approximate formulation for unit commitment problem

doi:10.1080/23307706.2025.2563791

2025-10-01

PRE

Chen Zhang *

Linfeng Yang

DOI：10.1080/23307706.2025.2563791

原文链接

原文求助

摘要

En 中文

Unit commitment (UC) problem in the electricity market becomes more challenging due to the coupling ramping constraints it entails. This paper aims to address this challenge by introducing a discrete unit output formulation for the UC. Instead of using semi-continuous unit output variables, this formulation utilises a new binary variable to represent the discrete generation output of each unit. It reduces the search space for feasible regions. Furthermore, this new discrete unit output binary variable allows for the establishment of a compact and efficient constraint for the unit, which is commonly used to obtain new 2-period compactness ramping constraints, tighter multi-period ramping constraints, upper/lower limit constraints for unit output within the minimum up/down time, and cold/hot start constraints for the unit. These valid inequalities are then utilised to reformulate a novel binary integer linear programming traceable approximate formulation of the UC. To evaluate the performance of the discrete unit output formulations, they are compared with three other commonly used UC models using 62 instances. The results show that the proposed approach leads to improved solution times, particularly in longer time-period instances. The case study results validate the accuracy and effectiveness of the approach, which also offers superior computational cost reduction.

Keyword:

Unit commitment

binary integer linear programming

tight

compact