Page 31 -

24 P. Davidsson and H. Verhagen

computer simulation and, in particular, computer simulation of social complexity,
which concerns the imitation of the behaviour of one or more groups of social
entities and their interaction.
Computer simulation, as any other computer programme, can be seen as a tool,
which could be used professionally or used in the user’s spare time, e.g. when
playing computer games. It is possible to distinguish between different types of
professional users, e.g. scientists who use simulation in the research process to
gain new knowledge, policy-makers who use it for making strategic decisions,
managers (of a system) who use it to make operational decisions, and engineers
who use it when developing systems. We can also differentiate two user situations,
namely, the user as participant in the simulation and the user as observer of the
simulation. Computer games and training settings are examples of the former, where
the user is immerged in the simulation. In the case of using simulation as a tool for,
say, scientific research or decision support, the user is an outside observer of the
simulation. (In other words, we may characterize this difference as that between
interactive simulations and batch simulations.)
The main task of computer simulation is the creation and execution of a formal
model of the behaviour and interaction (of the entities) of the system being
simulated. In scientific research, computer simulation is a research methodology
1
that can be contrasted to empirically driven research. As such, simulation belongs
to the same family of research as analytical models. One way of formally modelling
a system is to use a mathematical model and then attempt to find analytical solutions
enabling the prediction of the system’s behaviour from a set of parameters and
initial conditions. Computer simulation, on the other hand, is often used when
simple closed form analytic solutions are not possible. Although there are many
different types of computer simulation, they typically attempt to generate a sample
of representative scenarios for a model in which a complete enumeration of all
possible states would be prohibitive or impossible.
It is possible to make a general distinction between two ways of modelling the
system to be simulated. One is to use mathematical models and is referred to as
equation-based (or system dynamics or macro-level) simulation. In such models,
the set of individuals (the population of the system) is viewed as a structure that can
be characterized by a number of variables. In the other way of modelling, which
is referred to as individual-based (or agent-based or micro-level) simulation, the
specific behaviours of specific individuals are explicitly modelled. In contrast to
equation-based simulation, the structure is viewed as emergent from the interactions
between the individuals, thus exploring the standpoint that complex effects need not
have complex causes. We will here, as well as in the remainder of this book, focus
on individual-based simulation.

1
This distinction is of course not set in stone. For an example of an evidence-driven approach to
computer simulation, see Chap. 27 in this volume (Geller and Moss 2017).