Computational economics is an advanced field of research in which economists use computational tools to solve analytical problems and predict or model the complex interaction of many agents within financial marketplaces. In a nonlinear, dynamic system, such as the financial market, computational economics finds numerical solutions to multifaceted problems in the areas of public finance, game theory, and macroeconomics. Through the construction of virtual economics systems, researchers can broaden their understanding of observed regularities, norms, and potential behaviors within these systems in the real world. On the other hand, formulas and algorithms that are derived purely from economic reasoning rarely correlate with reality. Many computational economics software packages with pre-programmed optimization routines, such as Gauss or Conopt, allow economists to harness expanding computer capabilities to generate solutions to economics problems.
Agent-based computational economics (ACE) is a specialized area of computational economics that constructs models of economic processes using fixed initial conditions determined by the researcher but thereafter developed based on the interaction of a variety of independent factors or agents. Agents may be physical entities, such as the weather, or biological entities, such as livestock or crops. Additionally, individuals, such as consumers, or institutions, such as a major market, may also constitute agents. Researchers select a particular set of agents for a given formulation based on the problem being studied.
Several critical goals drive computational economics. Some researchers seek to evaluate the system performance of given processes, organizations, or policies virtually before they are instituted to make sure that the outcomes are socially advantageous. They introduce a variety of agents, such as consumers, producers, regulatory agencies, and other pertinent factors, all with private motivations and learning capacities, and allow those agents to interact virtually and develop over time. The main question that the researchers try to answer is whether strategic behavior by any of the agents may produce unfair advantage, inefficiency, or disorder in the system. In this way, researchers may identify deficiencies in a policy, plan, or system before implementation.
Another goal of computational economics is the identification of regularities and patterns within an economic system, which at first glance may seem to be totally random. Researchers attempt to explain the observed regularities by tracking recurrent patterns of interaction among the agents. Using computerized simulations, economists have been able to make out recurring patterns in business cycles, market procedures, and trade networks. Advances in methodology in programming, statistical analysis and visualization have expanded the capacity of computational economists to understand economic phenomena and generate theories regarding what causes them.