Simulation & Gaming

We can enhance discussions of simulation by distinguishing between the rationale for using simulation and the philosophical tools for conceptualizing it. I offer my thoughts, inspired by this symposium, on the proposition that simulation is little more than stuff of our minds, a figment of our imagination. I also argue that philosophy is important to all simulation/gaming work.

Philosophy finds simulation and gaming intriguing because of the epistemological issues that they raise. For example, how do simulations of a phenomenon help to explain that phenomenon? This collection of essays addresses such issues and hopes to stimulate additional investigations in this fascinating area.

Some computer simulations have explanatory goals. In a typical case, the computer simulation exemplifies a model generating a phenomenon under investigation. For the simulation to be explanatory, the model has to be explanatory. Models that fully explain phenomena are rare, but a model may offer a partial explanation of a phenomenon. It does this if an isomorphism holds between certain features of the model and certain features of a natural system realizing the phenomenon. The first two sections elaborate this account of a simulation’s explanatory power. The third section illustrates the account by applying it to Brian Skyrms’s (2004) simulations of the evolution of cooperation. The final section uses the account to suggest ways of increasing those simulations’ explanatory power.

Computer simulations of complex physical objects and processes for which data are very sparse or inexistent have become a major tool of scientific investigation in astrophysics and cosmology. However, one must ask how these simulations acquire their epistemic credentials and whether their realistic ambition is legitimate. A close look at two model-building processes—one in galactic astrophysics, the other in cosmology—reveals heretofore underappreciated features of simulations, such as path dependency. This article argues that such features undermine our confidence in the outcomes of the simulation. Case studies presented here reveal a general tension in computer simulation between realistic ambitions and the possibility of empirical confirmation. The analysis will thus lead to a reassessment of the epistemic goals actually achieved by composite models of complex astrophysical and cosmological phenomena.

This is an investigation of the relationship between simulations and models, concluding that the model intended to underlie a simulation and the model that actually underlies the simulation can be quite different. It is also argued that simulations and models are complexes of representational relations, making it the case that determining what is a model or what is a simulation is an inherently observer relative task that can never be fully overcome. The end result is that simulations can still be profitably used as scientific tools, but that caution must be used in interpreting their results; and we must be reluctant to make hard and fast judgments regarding what counts as a simulation or model and how any results obtained should be used.

Meteorology has employed automatic computing machines since the early days of electronic computers. From the 1950s on, a large body of models used for “in silico” experiments (numerical simulation) has been built up, together with an international infrastructure of measuring, modeling, and testing. These outstanding developments— unique in science—led not only to an increasing standardization in developing and applying models but also to deepening the interlinking between modeling and generating evidence. The article explores needs and strategies for evaluating scientific results based on mass data output devices.

Agent-based simulation (ABS) studies have recently been employed to support policy decisions. This article addresses the particular potentials and problems that ABS faces in this usage. First, the author warns against taking “familiarity” with specific ABS as a criterion for having confidence in the model’s policy recommendations. Second, he shows that specific epistemic issues—in particular the high number of detailed simulated systems—require additional reflection on which decision rules to choose for policy decisions based on ABS. Third, the author points out directions in which the construction and uses of ABS in policy decision could be improved. Each of these issues is illustrated by simulation studies undertaken to investigate smallpox vaccination policies.

In this article, the author argues that simulation is an undervalued technique to draw conclusions about empirical phenomena in economics. If the aim is to learn about the behavior of socioeconomic systems of interest, simulations have a variety of advantages relative to alternatives such as mathematical (pen and paper) modeling and laboratory experimentation. Therefore, the author has a good prima facie reason to exploit this method more fully. The author proceed by demonstrating that frequently heard arguments against simulations are wrong, and finally the author discusses a number of more specific empirical phenomena, criticisms of one type of simulation methodology used in economics.

Simulations facilitate the construction of increasingly complex theories, which, in turn, suffer three attendant curses. By definition, complex theories have many variables, decreasing the chances that all of them will be clear and measurable. This makes complex theories difficult to test, decreasing the chances they will be adequately tested, and difficult to understand, and decreasing the chances of attracting or sustaining an audience. However, constructing simulations that embody complex theories provides an opportunity for new perspectives and insights, and an occasion to renew respect for the complexities of the phenomena that they simulate. These are aesthetic benefits concordant with the definition of science as art with numbers.