Electricity market analysis using implicit learning models

Location of Research: Remote


The student will use a new tool set (http://implicit-layers-tutorial.org/) on machine learning using implicit layers to analyze electricity markets. The student will first establish a daily market model including generator cost and constraints, but exact parameters in this model are unknown. Then the student will perform training using historical price, demand, and renewable data and identify unknown parameters in the model. Eventually the student will test the developed model using data out of the training set to demonstrate performance.

Lab: Xu's lab

Direct Supervisor: Bolun Xu

Position Dates: 5/15/2021 - 8/15/2021

Hours per Week: 20

Paid Position: Yes

Credit: Yes

Qualifications: The student should be familiar with Python and standard machine learning packages (Tensorflow, PyTorch, CVXPY). Experiences with power system and electricity markets are a plus.

Eligibility: Junior, Senior; SEAS only