Project

Back to overview

Foundations of Market User Interface Design and Applications to the Smart Grid

English title Foundations of Market User Interface Design and Applications to the Smart Grid
Applicant Seuken Sven
Number 147211
Funding scheme Project funding (Div. I-III)
Research institution Institut für Informatik Universität Zürich
Institution of higher education University of Zurich - ZH
Main discipline Information Technology
Start/End 01.04.2013 - 31.05.2017
Approved amount 179'600.00
Show all

All Disciplines (2)

Discipline
Information Technology
Economics

Keywords (12)

User Interface Design; Smart Grid; Market Design; Behavioral Economics; Decision Making; Market User Interfaces; Hidden Markets; Preference Elicitation; Artificial Intelligence; Machine Learning; Active Learning; Electronic Markets

Lay Summary (German)

Lead
In immer mehr Bereichen unseres Lebens interagieren wir mit Märkten. Aber angesichts der Vielzahl dieser Märkte ist es für die Endnutzer unmöglich, jederzeit optimal zu handeln. In diesem Forschungsprojekt beschäftigen wir uns daher mit der Frage, wie man bessere Markt-Nutzer-Schnittstellen entwickeln kann, so dass die Interaktion für den Nutzer vereinfacht bzw. automatisiert wird. Wir konzentrieren uns dabei insbesondere auf die Entwicklung neuer Markt-Nutzer-Schnittstellen für das Smart Grid.
Lay summary

Neue elektronische Märkte führen oft zur Erhöhung des gesellschaftlichen Wohlstandes, aber die Vielzahl der Märkte stellt die Endnutzer auch vor neue Herausforderungen. Im Laufe eines einzigen Tages kaufen wir einen Nachrichten-Artikel im Internet, eine neue App auf dem Smartphone, und einen Film auf iTunes. Bei so vielen Markt-Interaktionen ist es schier unmöglich, immer eine optimale Markt-Entscheidung zu treffen.

Das Ziel unserer Forschung ist es, einfachere/automatisierte Markt-Nutzer-Schnittstellen zu entwerfen, die den Gesamtnutzen der Endnutzer maximieren. Um dieses Ziel zu erreichen, verfolgen wir zwei Ansätze. Zum einen werden wir untersuchen, wie Nutzer-Schnittstellen aussehen sollten, so dass die Endnutzer in der Lage sind, ökonomisch gute Entscheidungen zu treffen. Zum anderen werden wir neue Algorithmen entwickeln, die in der Lage sind, die Präferenzen der Endnutzer zu lernen, so dass ein Teil der Markt-Interaktion automatisiert werden kann. Dies ist von besonderer Wichtigkeit für die Realisierung der Smart Grid Vision. Um die Energieproduktion und den Energieverbrauch besser ausbalancieren zu können, wäre es ideal, wenn der Strompreis auch für die Endverbraucher innerhalb eines Tages dynamisch angepasst werden könnte. Allerdings wäre es einem Endnutzer natürlich unmöglich, zu jeder Sekunde optimal auf den dynamischen Strompreis zu reagieren. Deshalb sind hier halb-automatisierte Markt-Nutzer-Schnittstellen (mit Hilfe von Lern- und Planungs-Algorithmen) von besonderem Wert, sowohl für den einzelnen Endnutzer als auch für die Stabilität des Stromnetzes insgesamt.

Unsere Arbeit wird grundlegende Beiträge zur wissenschaftlichen Forschung in den Bereichen Markt-Design, Nutzer-Schnittstellen-Entwurf, und Künstliche Intelligenz leisten. Darüber hinaus erwarten wir, dass unsere Forschung einen Einfluss auf die Entwicklung im Smart Grid Bereich haben wird, und dass wir dadurch helfen können, die Realisierung der Smart Grid Vision voran zu treiben.

Direct link to Lay Summary Last update: 05.08.2013

Lay Summary (English)

Lead
In the next decade, we will encounter an abundance of new electronic markets in our lives. But given the multitude of markets, it is impossible for end-users to always make optimal economic decisions in these markets. In this research project, we study how to design better “market user interfaces” such that the interaction for the user is simplified or even automated. In particular, we focus on the design of market user interfaces for the smart grid (i.e., the next generation of the power grid).
Lay summary

New electronic markets often lead to social welfare improvements, but they also lead to new challenges for the end-users. During a single day, we may buy a news article online, an app on our smartphone, and a movie on iTunes. Given the multitude of markets in our lives, it is impossible to always make an optimal economic decision.

The goal of our research project is to design simpler (or automated) “market user interfaces” such that the total utility of the end-users is maximized. To reach this goal, we will pursue two different approaches. First, we will study how market user interfaces should be designed such that end-users are able to make good economic decisions. Second, we will develop new algorithms that are able to learn users’ preferences over time, such that part of the market-user interaction can be automated. This automation is particularly important for the realization of the smart grid vision. To better balance energy production and consumption, it would be ideal if end-users would also be exposed to dynamic energy prices. However, it would obviously be impossible for end-users to optimally react to energy prices that are changing every second. For this reason, half-automated market user interfaces (using learning and planning algorithms) are particularly valuable in this domain, for the individual end-user as well as for the stability of the power grid as a whole.

Our work will make significant contributions to the academic research in the areas of market design, user interface design, and artificial intelligence. Furthermore, we expect that our research will have a positive impact on the development efforts in the smart grid domain, and we believe that our project can help make the smart grid vision become a reality.

Direct link to Lay Summary Last update: 05.08.2013

Responsible applicant and co-applicants

Employees

Publications

Publication
Save Money or Feel Cozy? A Field Experiment Evaluation of a Smart Thermostat that Learns Heating Preferences
Shann Mike, Alan Alper, Seuken Sven, Costanza Enrico, Ramchurn Sarvapali (2017), Save Money or Feel Cozy? A Field Experiment Evaluation of a Smart Thermostat that Learns Heating Preferences, in Proceedings of the 16th International Conference on Autonomous Agents and Multiagent Systems (AAMAS), International Foundation for Autonomous Agents and Multiagent System, New York, NY.
It is Too Hot: An In-Situ Study of Three Designs for Heating
Alan Alper T., Shann Mike, Costanza Enrico, Ramchurn Sarvapali D., Seuken Sven (2016), It is Too Hot: An In-Situ Study of Three Designs for Heating, in Proceedings of the 2016 SIGCHI Conference on Human Factors in Computing Systems, 5262-5273, ACM, New York, NY5262-5273.
Adaptive Home Heating under Weather and Price Uncertainty using GPs and MDPs
Shann Mike, Seuken Sven (2014), Adaptive Home Heating under Weather and Price Uncertainty using GPs and MDPs, in Proceedings of the International Conference on Autonomous Agents and Multiagent Systems (AAMAS), International Foundation for Autonomous Agents and Multiagent System, New York, NY.
An Active Learning Approach to Home Heating in the Smart Grid
Shann Mike, Seuken Sven (2013), An Active Learning Approach to Home Heating in the Smart Grid, in Proceedings of the Twenty-Third International Joint Conference on Artificial Intelligence, 2892-2899, AAAI Press, Palo Alto, CA2892-2899.

Collaboration

Group / person Country
Types of collaboration
Prof. Dr. Jens Strueker (Fresenius University, Germany) Germany (Europe)
- in-depth/constructive exchanges on approaches, methods or results
Dr. Sarvapali Ramchurn (University of Southampton) Great Britain and Northern Ireland (Europe)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
- Exchange of personnel
Enrico Costanza (University College London) Great Britain and Northern Ireland (Europe)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
- Exchange of personnel

Scientific events

Active participation

Title Type of contribution Title of article or contribution Date Place Persons involved
16th International Conference on Autonomous Agents and Multiagent Systems (AAMAS) Talk given at a conference Save Money or Feel Cozy? A Field Experiment Evaluation of a Smart Thermostat that Learns Heating Preferences 08.05.2017 Sao Paulo, Brazil Seuken Sven;
13th International Conference on Autonomous Agents and Multiagent Systems (AAMAS) Talk given at a conference Adaptive Home Heating under Weather and Price Uncertainty using GPs and MDPs 05.05.2014 Paris, France Seuken Sven; Shann Mike;
23rd International Joint Conference on Artificial Intelligence (IJCAI) Poster An Active Learning Approach to Home Heating in the Smart Grid 03.08.2013 Beijing, China Shann Mike;
First Annual Workshop on Crowdsourcing and Online Behavioral Experiments (COBE) at EC Talk given at a conference Towards Personalized Market User Interfaces: An Online Behavioral Experiment 17.06.2013 Philadelphia, United States of America Seuken Sven;


Knowledge transfer events

Active participation

Title Type of contribution Date Place Persons involved
Züri ML #16: Cloud Machine Learning and Home Heating Talk 24.06.2015 Zürich, Switzerland Shann Mike;


Communication with the public

Communication Title Media Place Year
Talks/events/exhibitions Präsentation im Rahmen der Informatiktage 2017 German-speaking Switzerland 2017

Abstract

In the next decade, we will encounter an abundance of new electronic markets in our lives. This has manifold advantages, as market-based systems are often able to allocate resources more efficiently than non-market based systems. Consider, for example, the use of dynamic toll roads, the use of dynamically priced energy, or the many ways in which we can consume content for small amounts of money on the Internet.In recent years, economists and computer scientists have significantly advanced our understanding of the design of such market-based systems, giving rise to a new field called "market design." This new research area captures many different markets, ranging from matching markets (e.g., school choice problem, hospital-doctor matching, etc.), over combinatorial spectrum auctions, to the design of sponsored search auctions. One common, yet debatable feature of the large majority of the work in this field, is the assumption that the players in these markets are perfectly rational agents. In practice, however, users have cognitive costs, cognitive biases, they have limited time and bounded computational resources for deliberation, in short: "real" humans make mistakes when making decisions in market-based environments.This research proposal addresses this discrepancy between the common economic models and the actual psychological decision-making processes. We focus, in particular, on the intersection of "market design" and "user interface design," a research area that has been largely ignored so far. We propose new methodologies to design markets user interfaces (UIs) that take users' cognitive costs into account and develop new technological tools that make up for the cognitive limitations. The same way that color-coded planes make the job of an air-traffic controller easier, it is our goal to design market UIs that make economic decision making easier and thereby improve overall social welfare.The proposal consists of three complementary parts: First, we study the "foundations of market user interface design": how does the way we display information and what choices we offer users affect their decision-making capabilities. The PI has proposed the design of market user interfaces as a new research area in 2011, and we propose the following three new contributions: 1) the development of a behavioral user model to accurately predict user behavior in complex market environments and the identification of the main market UI design levers that we can tweak to optimize market UIs; 2) new methods to automatically design personalized market UIs based on users' cognitive skills; and 3) a collaborative-filtering method to design personalized market UIs based on the information available in a large group of market users.In the second part of the project we study how to design "smart market UIs for the smart grid." A key element of this part is the design of new learning, planning, and control algorithms to automate the interaction between the smart grid and the user. We propose the following four contributions: 1) the design of semi-automated market UIs involving active/online learning; 2) the design of forward-looking market UIs for the smart grid that take into account predictable changes in the market price as well as endogenous changes in the environment; 3) the design of smart grid UIs for dynamic environments that automatically adapt to changes in exogenous factors; and 4) the design of smart grid UIs for non-stationary user preferences.In the third part of the project we will develop prototype market UIs for the smart grid. We propose the following two contributions: 1) the development of a prototype market UI for the home heating control problem, i.e., the semi-automated control of a thermostat in a user's house; and 2) the development of a prototype market UI for the electric vehicle charging problem, i.e., a UI that can elicit a user's preferences regarding his trade-off between a fully or partially charged electric car vs. small or large cost savings.Significance of the Research: The proposed research will lay the essential foundations of research at the intersection of market design and UI design necessary for further study of this area. We will significantly advance both the theory and practice of market UI design. We expect that our theoretical and experimental research will constitute the building blocks of future academic work on this topic. Furthermore, the techniques we develop for the smart grid market (learning, planning, and control algorithms) as well as the specific market UIs for the home heating and the electric vehicle charging problems, will have an immediate impact on research and development for the smart grid, defining the state of the art of market UI design in academic and industry.
-