Most economic and business systems are complex, dynamic, and nondeterministic systems. Different modeling techniques have been used for representing real life economic and business organizations either on a macro level (such as national economics) or micro level (such as business processes within a firm or strategies within an industry). Even though general computer simulation was used for modeling various systems (Zeigler, 1976) since the 1970s the limitation of computer resources did not allow for in-depth simulation of dynamic social phenomena. The dynamics of social systems and impact of the behavior of individual entities in social constructs were modeled using mathematical modeling or system dynamics. With the growing interest in multi agent systems that led to its standardization in the 1990s, multi agent systems were proposed for the use of modeling social systems (Gilbert & Conte, 1995). Multi agent simulation was able to provide a high level disintegration of the models and proper treatment of inhomogeneity and individualism of the agents, thus allowing for simulation of cooperation and competition. A number of simulation models were developed in the research of biological and ecological systems, such as models for testing the behavior and communication between social insects (bees and ants). Artificial systems for testing hypothesis about social order and norms, as well as ancient societies (Kohler, Gumerman, & Reynolds, 2005) were also simulated. Since then, agent-based modeling and simulation (ABMS) established itself as an attractive modeling technique (Klugl, 2001; Moss & Davidsson, 2001). Numerous software toolkits were released, such as Swarm, Repast, MASON and SeSAm. These toolkits make agent-based modeling easy enough to be attractive to practitioners from a variety of subject areas dealing with social interactions. They make agent-based modeling accessible to a large number of analysts with less programming experience.
Computer simulation modeling is an established method in scientific and industrial applications, appropriate for obtaining insight into the dynamics of organizations. Modeling is used to represent a part of reality in sufficient detail, and resulting model is an artificial system used for experimentation. There are several situations when replacement of the real system by an artificial one is helpful or even necessary.
Inaccessibility of the real world system: Sometimes a part of the real world system that should be studied, is not accessible either because the system does not exist any more or is not yet put into operation.
Real world system is inappropriate for experimentation: Some real world systems may be affected in undesired way by experimentation. Examining effects of drastic changes in taxing and pricing policies may for example, disturb the fiscal system, or discourage production and consumption.
Time scale or behavior of the system is inappropriate for observation: A number of systems such as investments in some industries generate results over long periods of time, making it hard to collect enough data from the real system for a meaningful analysis. Simulation is using virtual time that can be accelerated or slowed down as needed in order to observe a particular phenomenon.
Intensive dynamics of the system: All elements of simulation model can be taken under full control. This is especially important in economics for the purpose of studying the impact of changes in one factor on behavior of the whole system, while holding all other factors at the same level. This presumption cannot be achieved in a real life economic system (e.g., system of supply and demand).
Key Terms in this Chapter
Artificial Environment: A model of the environment where the simulation model is operating. Environment model is completely controllable by the modeler. Particular environment models are highly relevant when modeling adaptive elements of the system or when using adaptive capabilities of the agents contained within the model.
Intelligent Software Agent: A system situated within a part of the environment that senses this environment and acts on it over time in pursuit of its own agenda.
Simulation Modeling: An established method in science and industry used to map a part of reality in sufficient detail using a model. Developed model should be able to answer questions directed to the real system without disturbing functioning of the real system.
Virtual Time: Denotes a time advancement paradigm used to handle the course of events within the simulation model. Event-based simulations use event queues that allow the simulation time to advance to the time stamp of the next event. In this way, the time scale can be stretched or compressed, depending on the needs of the model.
System Dynamics: A continuous simulation of systems exhibiting feedback loops. The feedbacks can either intensify activities of the system (positive feedback) or slow them down and stabilize the system (negative feedback).
Experimental Frame: Establishes the set of experiments for which the model is valid. It has to be determined in early stages of model development.
ABMS: Abbreviation for agent-based modeling and simulation; a synonym for multi agent simulation.
Multi Agent System: A system consisting of a number of agents that interact with each other through communication, thus allowing them to achieve goals that are beyond their individual capabilities.
Discrete-Event Simulation: A type of simulation where the simulation mechanism advances the simulation clock to discrete points in time. Time advancement can be round-based (simulation clock is advanced for a constant number of time units for each round of the simulation) or step-based (simulation clock is advanced to the time stamp of the next event in the event queue).
Multi Agent Simulation: A simulation modeling paradigm that uses software agents to represent the entities of the modeled system that interact with each other and with the virtual environment. These interactions are used to model dynamics in functioning and structure of the real system.