On the Solution Uniqueness of Data-Driven Modeling of Flexible Loads

Description

• main.m: start from this file.

• generate_data.m: This script generates (price, power) data by simulating a dynamic energy management system with adjustable loads and virtual batteries. The code calculates the optimal power consumption under different price signals.

• identification.m: This script identifies a physics-based generalized flexible model, including the forward problem modeling, the derivation of the concise form using KKT conditions, and the inverse optimization problem solution.

• plot_fig1.m: This script visualizes the convex hull of $\Gamma$ (i.e., $\rm Conv(\Gamma)$) and the set $\Pi$ under different numbers of samples.

• plot_fig2.m: This script plots the feasible region of the aggregated power of $\rm{Ω}_{\rm phy}$ under different numbers of virtual batteries.

Authors

• Jiayi Ding, School of Electrical Engineering, Southeast University, Nanjing, China
• Shuai Lu, School of Electrical Engineering, Southeast University, Nanjing, China

Usage

• Start from main.m

Dependencies

• MATLAB
• YALMIP toolbox
• Gurobi solver
• Multi-Parametric Toolbox 3 (MPT3)

Note

• To reproduce the results presented in an associated paper, set filename to 'random_price_benchmark.mat' in "generate_data.m".