Each student began a research project during the two-week workshop. Below are brief descriptions and links to these various projects. The expectation is that these projects will form the basis for dissertation chapters and/or journal articles.
Given the ambitious nature of the projects, this page is likely to change over the next few months as projects reach completion.
Todd is investigating the effectiveness of monitoring for resolving social dilemmas. Agents, for example, employees in a worker-managed firm, are allowed to monitor the effort level of other agents, and payoffs to all agents are based on these observations. Agents adaptively modify their monitoring strategies, and questions concerning the emergence of cooperation in such situations are analyzed.
Tom has created a general "artificial worlds" engine to explore issues of cooperation and altruism. Using this engine he is currently exploring issues of selective interaction among cooperative agents. In this model, agents alter their desired level of global interaction as well as direct connections with other agents through a process of learning and mutation.
Kim-Sau is exploring how systems react to ambiguous news about the state of a system. For example, the impact of a coup on a country's economy or, say, the release of unanticipated earnings news are likely to signal a large change in the environment, yet the actual magnitude of this change may be ambiguous. Using a dynamic mathematical model, Kim-Sau models both the waiting and learning processes inherent in such events.
Garett has developed a simple platform for investigating credit allocation mechanisms. This project is motivated by a larger on-going research program investigating artificial adaptive negotiation agents which are driven by sets of possibly incomplete and inconsistent rules-of-thumb in a classifier system design. Unfortunately, classifier systems are notoriously poor performers due in part to difficulties with the bucket brigade credit allocation mechanism. This project's goal is to investigate and compare the isolated effects of various credit allocation mechanisms on a very simple (but scalable) problem.
Jorge is analyzing different models of foreign exchange rate bands through Monte Carlo methods. This analysis will allow him to clarify the links between the theoretical models and the empirical findings. By constructing a FX model laboratory, he can begin to explore some normative issues across the models, for example, adaptive models of bank intervention policies can be analyzed.
Carl has begun to explore issues of learning in overlapping generation models using simple ecologies of agents. This work complements the existing work of learning in these models, and provides a means by which to generate a number of new and test a variety of existing hypotheses. His current focus is on competition among forecasting rules of various degrees of sophistication.
John is focusing on models of party competition where citizens abstain from voting when either party platforms are too far away from the voter's ideal point or when parties have closely matched (in terms of voter's utility) platforms. He finds that such abstention dramatically alters the form of the electoral landscape, and may lead to multiple equilibria and unusual adaptive dynamics.
Konstantina and Benedikt have created a model of industry competition using the Swarm platform. Their model incorporates ideas about learning-by-doing, entry and exit dynamics, and product differentiation in attribute space.
Scott is studying models of adaptive party competition. His initial work has focused on allowing voters to pay attention to various aspects of each party's platform. Voters are allowed to adaptively choose which issues to focus on, while simultaneously parties alter their platforms to capture votes. Current extensions include allowing parties to focus voter attention on particular issues through political "advertising."
Mariana is exploring some fundamental issues in industrial organization. Using a reduced form model of market share dynamics based on a replicator equation, she is analyzing the dynamics implied by various cost conditions. Along with the computational exploration of the model, she is also developing two complementary mathematical approaches.
Martin is analyzing learning in repeated games, through the use of an adaptive aspiration model. Agents maintain their past action in the game as long as their aspirations are met, however, when payoffs fall below aspiration levels they alter their play. Initial work on the repeated Prisoner's dilemma indicates that the above model results in some interesting dynamics.
Thor is investigating trade (based on a model by Diamond) under adaptive agents relying on finite automata rules. Such a model has many interesting implications vis-a-vis traditional methods. The model also offers a variety of new analytic possibilities, for example, the analysis of strategic complexity.
Matt is probing the dynamics of market structure and formation. In his model, agents search and establish relationships with different vendors, and in the process create a network of marker interactions.
Yury is studying the dynamics of spatial infrastructure formation. He has developed a model in which the self-organization of spatial infrastructure can be analyzed.
Eduardo is finding connections between the "El Farol" problem and game theoretic results on learning in games. Using these connections, he has begun to provide a variety of new theoretical insights into the basic problem.
Jose has created a general equilibrium model where self-interested market-makers drive the formation of new markets.