Title: Gas networks analytics with constraints
Authors: Ying Chen - National University of Singapore (Singapore)
Nazgul Zakiyeva - National University of Singapore (Singapore) [presenting]
Thorsten Koch - Zuse Institute Berlin (Germany)
Bangzhu Zhu - Wuyi University (China)
Abstract: A Gas networks Analytics with Constraints (GAC) model is proposed for large scale network, where supply and demand in the network are balanced. Simultaneously considering three challenges in the GAC framework, namely 1) high dimensionality and 2) unknown adjacency matrix and 3) equality constraints, we estimate the constrained adjacency matrix under sparsity assumption. After discussing the estimation procedures and theoretical properties of the GAC model, we demonstrate its forecasting performance using real life network data and compare with alternative high dimensional models.