ABSTRACT. Many economic models have, to ensure tractability, excluded qualitative and complex dynamic phenomena, focussing instead on static equilibria. This affects how accurately they reflect reality. A combination of fuzzy logic and genetic algorithms should make more complex models tractable, by allowing agents to learn their own rules for surviving and prospering. This combination was used to examine if and how a market, with firms initially trading at different prices, can reach equilibrium (problem of the Walrasian Crier). Under perfect competition, the model finds the normal static equilibrium price; under more realistic conditions, with nonlinearities and risk, it reaches a quantitatively different equilibrium. (pp. 55–78) JEL Classification: C63, E13, P51