CAEPIA 2011: Advances in Artificial Intelligence pp 32-41 | Cite as
Market Self-organization under Limited Information
Abstract
The process of gradually finding an economic equilibrium, the so called tâtonnement process, is investigated in this paper. In constrast to classical general equilibrium modelling, where a central institution with perfect information about consumer preferences and production technologies (“Walrasian auctioneer”) organizes the economy, we simulate this process with learning consumer and producer agents, but no auctioneer. These agents lack perfect information on consumption preferences and are unable to explicitly optimize utility and profits. Rather, consumers base their consumption decision on past experience – formalized by reinforcement learning – whereas producers do regression learning to estimate aggregate consumer demand for profit maximization. Our results suggest that, even without perfect information or explicit optimization, it is possible for the economy to converge towards the analytically optimal state.
Keywords
Agent-based computational economics market self-organisation reinforcement learning regression learningPreview
Unable to display preview. Download preview PDF.
References
- 1.Bowling, M., Veloso, M.: Multiagent learning using a variable learning rate. Artif. Intell. 136(2), 215–250 (2002)Google Scholar
- 2.Busoniu, L., Babuska, R., De Schutter, B.: A comprehensive survey of multiagent reinforcement learning. IEEE T. Syst. Man. Cy. C. 38(2), 156–172 (2008)Google Scholar
- 3.Colander, D., Rothschild, C.: Sins of the sons of samuelson: Vision, pedagogy, and the zig-zag windings of complex dynamics. J. Econ. Behav. Organ. 74(3), 277–290 (2010)Google Scholar
- 4.Erev, I., Roth, A.E.: Predicting how people play games: Reinforcement learning in experimental games with unique, mixed strategy equilibria. Am. Econ. Rev. 88(4), 848–881 (1998)Google Scholar
- 5.Gintis, H.: The dynamics of general equilibrium. Econ. J. 117(523), 1280–1309 (2007)Google Scholar
- 6.Gintis, H.: The dynamics of generalized market exchange. Santa Fe Institute Working Paper (2011)Google Scholar
- 7.Hsieh, C.Y., Mangum, S.L.: A search for synthesis in economic theory. M.E. Sharpe, New York (1986)Google Scholar
- 8.Kirman, A., Vriend, N.: Evolving market structure: An ace model of price dispersion and loyalty. J. Econ. Dyn. Control 25(3-4), 459–502 (2001)Google Scholar
- 9.Rieskamp, J., Busemeyer, J., Laine, T.: How do people learn to allocate resources? comparing two learning theories. J. Exp. Psychol. Learn. 29(6), 1066–1081 (2003)Google Scholar
- 10.Rouchier, J.: Agent-based simulation as a useful tool for the study of markets. Technical Report 8, GREQAM (2008)Google Scholar
- 11.Shoham, Y., Powers, R., Grenager, T.: If multi-agent learning is the answer, what is the question? Artif. Intell. 171(7), 365–377 (2007)Google Scholar
- 12.Spall, J.: Introduction to stochastic search and optimization: estimation, simulation, and control. Wiley Interscience, Hoboken (2003)Google Scholar
- 13.Tesfatsion, L., Judd, K.L.: Handbook of Computational Economics, vol. 2: Agent-Based Computational Economics. North-Holland, Amsterdam (2006)Google Scholar
- 14.Walras, L.: Elements of pure economics, or, The theory of social wealth. Allen and Unwin, London (1874/1954)Google Scholar
- 15.Wang, X., Sandholm, T.: Reinforcement learning to play an optimal nash equilibrium in team markov games. In: Becker, S., Thrun, S., Obermayer, K. (eds.) Advances in Neural Information Processing Systems, vol. 15, pp. 1571–1578. MIT Press, Cambridge (2003)Google Scholar
- 16.Weisbuch, G., Kirman, A., Herreiner, D.: Market organisation and trading relationships. Econ. J. 110(463), 411–436 (2000)Google Scholar