The focus of this study is to build, from the `bottom-up´, a market with artificiallyintelligent adaptive agents based on the institutional arrangement of the ColombianForeign Exchange Market (1994-1999) in order to determine simple agents´ design,rules and interactions that are sufficient to create interesting behaviours at themacroscopic level - emerging patterns that replicate the properties of the time seriesfrom the case study.Tools from artificial intelligence research, such as genetic algorithms and fuzzy logic,are the basis of the agents´ mental models, which in turn are used for forecasting,quoting and learning purposes in a double auction market. Sets of fuzzy logic rules yieldadequate, approximately continuous risk and utility preferences without the need to fixtheir mathematical form ex-ante.Statistical properties of financial time series are generated by the artificial market, aswell as some additional non-linearity linked to the existence of a crawling band.Moreover, the behaviour of the simulated exchange rate is consistent with currencyband theory.Agent´s learning favours forecasting rules based on regulatory signals against rulesbased on fundamental information. Also, intra-day volatility is strongly linked to therate of arrival and size of real sector trades. Intra-day volatility is also a function of thefrequency of learning and search specialisation. It is found that when a moderately lowfrequency of learning is used, volatility increases.
