Abstract: Agent-based computational economics (ACE) is the computational study of economies modeled as evolving systems of autonomous interacting agents. Thus, ACE is a specialization to economics of the basic complex adaptive systems paradigm. This study outlines the main objectives and defining characteristics of the ACE methodology, and discusses similarities and distinctions between ACE and artificial life research. Eight ACE research areas are identified, and a number of publications in each area are highlighted for concrete illustration. Open questions and directions for future ACE research are also considered. The study concludes with a discussion of the potential benefits associated with ACE modeling, as well some potential difficulties. Keywords: Agent-based computational economics; artificial life; learning; evolution of norms; markets; networks; parallel experiments with humans and computational agents; computational laboratories. 1

This paper investigates the dynamics in a simple present discounted value asset pricing model with heterogeneous beliefs. Agents choose from a finite set of predictors of future prices of a risky asset and revise their `beliefs' in each period in a boundedly rational way, according to a `fitness measure' such as past realized profits. Price fluctuations are thus driven by an evolutionary dynamics between different expectation schemes (`rational animal spirits'). Using a mixture of local bifurcation theory and numerical methods, we investigate possible bifurcation routes to complicated asset price dynamics. In particular, we present numerical evidence of strange, chaotic attractors when the intensity of choice to switch prediction strategies is high.

This paper presents results from an experimental computer simulated stock market. In this market artificial intelligence algorithms take on the role of traders. They make predictions about the future, and buy and sell stock as indicated by their expectations of future risk and return. Prices are set endogenously to clear the market. Time series from this market are analyzed from the standpoint of well-known empirical features in real markets. The simulated market is able to replicate several of these phenomenon, including fundamental and technical predictability, volatility persistence, and leptokurtosis. Moreover, agent behavior is shown to be consistent with these features, in that they condition on the variables that are found to be significant in the time series tests. Agents are also able to collectively learn a homogeneous rational expectations equilibrium for certain parameters giving both time series and individual forecast values

The use of computer simulated markets with individual adaptive agents in finance is a new, but growing field. This paper explores some of the early works in the area concentrating on a set of some of the earliest papers. Six papers are summarized in detail, along with references to many other pieces of this wide ranging research area. It also covers many of the questions that new researchers will face when getting into the field, and hopefully can serve as a kind of minitutorial for those interested in getting

Abstract not found

Abstract not found

The use of computer simulated markets with individual adaptive agents in finance is a new, but growing field. This paper explores some of the early work in the area concentrating on a set of some of the earliest papers. Six papers are summarized in detail, along with references to many other pieces of this wide ranging research area. It also covers many of the questions that new researchers will face when getting into the field, and hopefully can serve as a kind of minitutorial for those interested in getting started. ∗The author is grateful to the Alfred P. Sloan Foundation for support. Cars Hommes provided useful comments on an earlier draft. 1

The many motivations for employing agent-based computation in the social sciences are reviewed. It is argued that there exist three distinct uses of agent modeling techniques. One such use — the simplest — is conceptually quite close to traditional simulation in operations research. This use arises when equations can be formulated that completely describe a social process, and these equations are explicitly soluble, either analytically or numerically. In the former case, the agent model is merely a tool for presenting results, while in the latter it is a novel kind of Monte Carlo analysis. A second, more commonplace usage of computational agent models arises when mathematical models can be written down but not completely solved. In this case the agent-based model can shed significant light on the solution structure, illustrate dynamical properties of the model, serve to test the dependence of results on parameters and assumptions, and be a source of counter-examples. Finally, there are important classes of problems for which writing down equations is not a useful activity. In such circumstances, resort to agent-based computational models may be the only way available to explore such processes systematically, and constitute a third distinct usage of such models.

Abstract not found

In this paper we analyze a simple model of adaptive competition which captures essential features of a variety of adaptive competitive systems in the social and biological sciences. In this model, each of N agents, at each time step of a game, joins one of two groups. The agents in the minority group are awarded a point, while the agents in the majority group get nothing. Each agent has a fixed set of strategies drawn at the beginning of the game from a common pool, and chooses his current best-performing strategy to determine which group to join. We find that for a fixed N, the system exhibits a phase change as a function of the size of the common strategy pool from which the agents initially draw their strategies. For small pool sizes, the system is in an efficient market phase in which all information that can be used by the agents ' strategies is traded away, and no agent can accumulate more points than would an agent making random guesses. In addition, in this phase the commons suffer, and relatively few points are awarded to the agents in total. For large initial strategy pool sizes, the system is in an inefficient market phase, in which there is predictive information available to the agents ' strategies, and some agents can do better than random at accumulating points. In addition, in this phase, the total number of points awarded to the agents is greater than in a game in which all agents guess randomly, and so the commons do relatively well. At a critical size of the strategy pool marking the cross-over from the efficient market to the inefficient market phases, the commons do best. This critical size of the pool grows monotonically, and in a very simple way with N. The behavior of this system has features reminiscent of a spin-glass in statistical physics, with the small pool size phase being, in a certain sense, more glassy than the large pool phase. We argue that the structure we have elucidated has important implications for a range of phenomena in the social and biological sciences, as well as for the general study of adaptive, competitive systems. 12/3/97 a