Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.

  

Stochastic Games for Smart Grid Energy Management with Learning Paradigm

Student(en):

Betreuer:

Suli Zou, John Lygeros
Beschreibung:

The electric grid is are evolving into a smart grid system integrating renewable resources, electric vehicles and smart meters etc. Among these, electric vehicles (EVs) play a crucial role in the energy management in smart grids. As shown recently, incorporating EVs into the grid design can significantly save the cost, shift the demand peak and fill the valley. They can also provide some other services such as frequency regulating, reserving and stabilizing the perturbation of renewable generation. Considering the role of EVs when renewables inject electricity into the grid, it will bring much flexibility in the energy management. Therefore, managing uncertainties using EVs and properly controlling the generation and demand are some of the most important challenges in the design and analysis of smart grids.

In the game formulation, EVs can be regarded as active players optimizing individual objectives. However, in such cases, the payoff received by each EV is a random variable with the randomness coming from the internal uncertainties during charging/discharging processes and the forecast error of renewable generation. As a result, although each EV has knowledge about its own utility, it is quite uncertain about the randomness caused by the others’ decision and generated energy. In this regard, there is some evidence that in the real world, decision makers do not make decisions based on expected values of outcomes evaluated by actual probabilities, but rather based on their perception on the potential value of losses and gains associated with an outcome.

For this project, I will schedule 1-2 meetings per week. The project will feature 80% of theory and 20% of simulation. Some basic knowledge on optimization, game theory and machine learning and a good level of mathematical maturity are required. Validation of the designed algorithms will require coding skills (MATLAB or Python).

Weitere Informationen
Professor:

John Lygeros
Projektcharakteristik:

Typ:
Art der Arbeit: 80% of theory, 20% of simulation
Voraussetzungen:
Anzahl StudentInnen:
Status: open
Projektstart: Feb. 2018
Semester: Spring 2018