John O. Ledyard - US grants

0090616
Murray

This award is to California Institute of Technology to support the activity described below for 30 months. The proposal was submitted in response to the Partnerships for Innovation Program Solicitation (NSF 0082).

Partners
Partners for the partnership include California Institute of Technology; Art Center College of Design; State government through Business Technology Center incubator; private industry.

Proposed Activities
The activities for this award include creation of post-degree entrepreneurial fellowships with the goal of preparing students previously trained in science or design to adapt their skills to the development of commercial products in the start-up environment; training in entreperneurialship (business plan, develop engineering prototypes; financial sources, etc); industrial partner mentor program.

Proposed Innovation
The innovation goals for the award include education of entrepreneurial leaders who have primary graduate and post-graduate education in science and engineering; formation of start-up high tech companies; development of educational modules for entrepreneurial courses for export to other universities.

Potential Economic Impact
The potential economic outcomes include teaching modules for export to other schools; spin-off companies; network of entrepreneurs and industry partners; graduates in science and technology with entrepreneurial training for leadership roles in the private sector.

Potential Societal Impact
The major benefit to society from this award will be the creation of high tech jobs and an education methodology for training future leaders in an innovative society.

Mechanisms allow individuals with dispersed information to achieve desired group outcomes. Experiments confirm theoretical predictions that different parameter values and different mechanism rules can have important effects on the efficiency of outcomes and the speed of convergence. A computer-based behavioral model will be developed to test a wide range of mechanisms and institutions. The model is to be based on evolutionary algorithms. Each smart agent has a collection of different strategies to use in its decision-making processes. Strategies compete for the right to make agent decisions through a selection process based on an evaluation of each strategy's hypothetical and actual performance. The model will first be tuned with mechanisms for which there are significant experimental data. Examples include the Groves-Ledyard mechanism, the voluntary contribution mechanism for public goods, the double auction, and simultaneous trading in a set of parallel markets. Then the model will be used to make predictions about behavior under those mechanisms that have not already been tested in experiments with human subjects. This project will develop a new methodology for economics research that has potential to be better, faster, and cheaper than experimental economics.

We propose to develop a framework for designing incentives for electric power networks and their associated markets that provides robust power generation while rewarding efficiency and environmental friendliness. The key technical thrusts are:

1. Development of prototype economic mechanisms for buying and selling
power addressing non-steady state performance and incorporating
engineering considerations such as production efficiency and
environmental emissions.

2. Analysis and synthesis of information fusion and feedback control
mechanisms at the component, network, and market levels providing high
performance and robust operation in the presence of uncertainty and
faults.

3. Implementation of economics experiments to test engineering performance
and market volatility of representative power networks, using 20-30
human subjects and software agents interacting with a distributed
simulation of a large scale power system.

We will combine methods from control, computation and economics in a unified framework for market-based systems that is expected to be applicable to other critical infrastructure problems involving interconnected economic, information, and engineering systems.

We will also develop elements of a curriculum that will provide training to students in economics, computer science, and engineering. These curriculum activities will be integrated into a novel set of interdisciplinary courses that are being developed for the newly formed Social and Information Systems Laboratory at Caltech. These courses will provide necessary training for economic and information scientists who are needed to analyze, design, implement and operate large scale social and information systems. Participation of women and underrepresented minorities will be specifically targeted and pursued through the summer undergraduate research programs.

One fundamental goal of social science is to develop good models of human behavior. But, it is difficult to know when we have been successful -- there are many competing models and, using standard methods, it is difficult to compare them to each other. In this research we advance a methodology involving a computer-based Turing Tournament in which data from emulators (computer programs designed to ''act like human subjects'' in repeated 2 person games) and data from human subjects are compared by detectors (computer programs designed to separate human from emulator data). The winning detector is the one best able to sort humans from emulators. The winning emulator is the one best able to fool the best detector. Unlike other tournaments, this is not one program playing another to win a game. The idea, after all, is not to be better than humans but to be exactly like humans. One of the broad benefits from this activity will be the creation of a publicly available test-bed in which researchers can test their models of behavior against true human behavior, and in which researchers can test the efficacy of their methodologies to detect non-human behavior.

Financial markets have long been known to play a crucial role in societal re-allocation and diffusion of risk. Recently, financial markets have been observed contributing to social cognition as well. Information is transmitted, problem solving is influenced, and individual inference is affected. The mechanics by which financial markets contribute to social cognition are not well understood. Neoclassical economic theory assumes that market participants can rationally infer information from others through transaction prices. But the very rationality on which such inference is based should make market participants wary of trading. Unfortunately, if there is no trade, there are no transaction prices, and hence, nothing is revealed. Social cognition is impossible.

People trade -- in fact they trade a lot -- but we do not know why. Correlation analysis of order and trade flows and subsequent actions has not provided much insight. Nor have surveys helped much, suggesting that actions may be largely sub-conscious. If so, direct measurement of sub-conscious changes in perceived risk and reward may be a necessary first step towards resolving the trading puzzle and eventually understanding the role of asset markets in social cognition. Recently, scientists have discovered how changes in expected reward and risk induces specific responses in certain sub-cortical parts of the brain. The PIs plan to reverse this approach, exposing subjects to market activity while monitoring brain activity. The goal is to detect features in order and trade flows that trigger changes in perceived risk and reward as reflected in brain activity. The approach borrows from the neuroscience of vision, where scientists have successfully been able to identify the sources of changes in visual perception even in environments as complex as full-feature movies.

Extensive work over the last decade, done within theoretical computer science, has provided deep insights into the computability of market equilibria for various market models and utility functions, using the powerful tools of the modern theories of algorithmic design and algorithmic complexity. This follows up on a century-long work, within mathematical economics, on obtaining a mechanism that converges to equilibrium -- a goal that had to be eventually abandoned due to certain negative results on efficient computability of the equilibrium. The work in TCS was motivated in part by applications to markets on the Internet.

The current project will extend this work along several exciting directions. Recent work of the lead Principle Investigator (PI) on using complementary pivot algorithms, for obtaining usable algorithms for certain market models that are unlikely to have efficient algorithms in the usual sense of polynomial worst-case running time, opens up the possibility of extending this approach to broader classes of markets, in particular, markets with production. A major new challenge is to address dynamically evolving markets.

In terms of applications of markets, the team brings to this project a wealth of experience on electricity markets, gained from work done with researchers in computer science and control dynamics. The PIs plan on bringing their expertise in mechanism design to bear on the problems of integrating renewable energy sources into the smart grid and providing better approaches to the pricing and allocation of ancillary services to guarantee reliability and stability. Another new challenge is to extend general equilibrium theory, the undisputed crown jewel of mathematical economics, to the digital economy. The traditional notion of equilibrium is not applicable to digital goods -- once produced, an unbounded number of copies of such goods are available. The digital realm is very rich and is increasingly occupying a larger share of our economy. It is imperative, therefore, to achieve the same depth of understanding of pricing for digital goods as was obtained for conventional goods.

This project will provide algorithms and insights into the computational aspects of markets, including electricity markets and transactions on the Internet, thereby helping make their operation more efficient. Hence, it is expected to contribute to advances in science and engineering, as well as to promote economic prosperity.

From the internet to the electricity grid to the interconnections of banks, networks are a pervasive feature of our world today. Their ubiquity has had a fundamental impact on modern markets. Driven by advances in information technology, modern marketplaces are becoming ever more dynamic, complex, and interconnected. While markets used to be slow to evolve, have simple institutions governing trade, and make finding the "right" trading partner a daunting challenge; markets emerging today are typically complex platforms, with a range of dynamic mechanisms for facilitating matches amongst participants. This gives an unprecedented level of control over the design and operation of markets, which has led to their incorporation as increasingly crucial components of networked systems, e.g., dynamic real time pricing in cloud computing markets and automated trading in financial markets. The changes in markets have also fundamentally impacted information technology, especially in the context of networked systems -- it is increasingly the case that networked systems that used to be purely "engineered" are now being governed by a mixture of markets and engineering, e.g., electricity markets are managed through a combination of multiple markets at different timescales (day ahead, real-time, etc.) and fast-time scale control systems to ensure stability.

Thus, networks and markets are connected today more than ever, and so it is necessary to study them as a unified whole rather than as individual fields. In particular, the incorporation of markets into networked systems creates an unavoidable interaction between "the engineer" and "the economist" in terms of which aspects of the system are managed via control policies, protocols, etc., which aspects are managed via economic tools such as pricing, market design, etc., and how the two interact. This is made even more challenging by the fact that engineering choices have consequences for economic efficiency and vice versa.

The researchers in this grant seeks to develop theoretical foundations for the study of networked markets, and will focus on three general themes: (i) the impact of connections between markets, (ii) the impact of connections between participants of markets, and (iii) understanding and controlling contagion in networked markets. These themes cut across many of the challenges and opportunities surrounding networked markets, and a better understanding of these issues is crucial for each of the five target applications on which the research focuses: cloud computing markets, data markets, electricity markets, matching markets (e.g. labor markets), and financial markets. In particular, the results of the research in this grant will provide new insights into the functioning of networked markets that can provide benefits for both market operators/designers and market participants.

The research is highly interdisciplinary, connecting computer science and economics and focusing on applications from diverse areas in both engineered and social systems. As part of the grant, the investigators are initiating a new PhD program on Computing and Mathematical Sciences at Caltech, which focuses training students with cross-cutting interdisciplinary interests at the interface of computing with science and engineering. Additional outreach activities include yearly workshops with both academic and industrial attendance. Further, the strong industry connections of the investigators in each of the application areas highlight the potential for the proposed research to make the transition from academia to practice.