Fully Adaptive Distributionally Robust Multi-stage Framework for Uncertain Unit Commitment Based on Mixed Decision Rules
================
What about this project/study?
With growing penetration of wind power into the power grid, while achieving low cost sustainable electricity supply, it also introduces technical challenges with the associated intermittency. This paper proposes a fully adaptive Wasserstein-based distributionally robust multi-stage frame-work based on mixed decision rules (MDR) for uncertain unit commitment problem (UUC) to better adapt wind pow-er respecting non-anticipativity both in unit state decision and dispatch process. Comparing with the existing multi-stage model, the proposed framework introduces an im-proved MDR to handle all decision variables to expand feasible region, thus, this framework can obtain various typical models by adjusting the number of relevant periods of decision variables. As a result, our model can find feasi-ble solution to some problems that are not feasible in the traditional models while finding better solution to feasible problems. The proposed model is reformulated with ad-vanced optimization method and improved MDR to form the mixed integer linear programming (MILP) model to address the computational intractability. The effectiveness and efficiency of the proposed model have been validated with case studies using IEEE benchmark systems.
The description of implement code files
BDR_UCC_DRO : Our proposed DR_MDR framework.
TS_UCC_DRO : The typical multi-stage DRO model.
ReadDataSCUC : Read the data.
ReadWindData : Read the wind data.
SCUC_nodeY : Construct network admittance matrix.
L_hat : Continuous lifting operator.
L_tine : Binary lifting operator.
Sumt : Calculate number of periods.
U_matrix : Construct vertices matrix.
Zbin : Construct binary coefficient vector.
Zboth : Construct mixed coefficient vector.
Zhat : Construct continuous coefficient vector.
Zhat_nonb : Construct continuous coefficient vector for typical models.
vers : Construct vertices set.
vers_nonb : Construct vertices set for typical models.
data: All IEEE datas in the simulations.
SCUC_X_Y ：IEEE X-bus Y-periods test system.
Wind_power : Historal wind data.
The current BDR_UCC_DRO is suitable for 6-bus test system.
If you want to expand it to a larger test system, you should revise the file path to corresponding system file,
and define decision variables for each buses, and align all variables with they decision variables.
Matlab R2018a
Cplex 12.7.1
Mosek 9.2
If you find our paper/code is useful to you, please cite it.
Authors：Ying Yang (yingyoung1997@gmail.com), Lingfeng Yang (ylf@gxu.edu.cn),Zhaoyang Dong
Team：www.scholat.com/team/eidp
Webpage: http://jians.gxu.edu.cn/default.do