Classical science of economics is based on the dualism between microeconomics, i.e. the study of the behavior of individuals and firms in the allocation of scarce resources, and macroeconomics, which studies the structure of the economy as a whole. In order to build a tractable bottom-up approach and provide for analytical results, rational behavior of economic entities is often assumed; diversity is ruled out by assumption, rationality is postulated, and macro-scale effects are obtained by scaling up the micro-level utility functions of representative agent forming a homogenous population of identical copies. Such an approach cannot capture the emergence of new phenomena from the interaction of heterogeneous interacting agents, the effects of irrational behavior in the society, or the dynamical behavior of the economy. 

Computer simulation, on the other hand, opens unprecedented possibilities for the exploration of complex economic systems, by means of interacting heterogeneous agents, behavior of which can be tracked in any level of detail in all dimensions of the problem formulation. The rigorous link from the behavioral micro-foundation to the macro-level aggregate characteristics can describe stable systems as well as transient behavior or extreme events such as crashes or bubbles. Reproducible research is possible by sharing simulation platforms and data sets, and emergence of new phenomena at all scales can be attributed to the actual cause. The aforementioned computational approach is further complemented by insights from statistical physics, theory of chaos, evolutionary algorithms, machine learning and other methods from the portfolio of computational techniques.