Modeling and Simulation

From Computing Concepts
Jump to: navigation, search

Models and simulations are simplified representations of more complex objects or phenomenon. For instance, student pilots use flight simulators to learn to fly. The simulator is a simplified (and safer) version of a full airplane.

Models may use different abstractions or levels of abstraction depending on the object or phenomenon being modeled. For instance tests of airplane components use models of strength and resilience of the physical components of the airplane (one form of abstraction of an airplane) for crash testing. Flight simulators model the actions and reactions of the airplane control system (a different abstraction of an airplane). Flight simulators may have different levels of abstraction that get closer and closer to controlling a real airplane. A simulator that runs using an actual airplane cockpit not as abstract as a simulation running on a personal computer. The personal computer version has abstracted away the actual equipment, but still preserves some of the actions and reactions of flight control – and is simpler.

Models and simulations can be used to test hypothesis that are formed to explain the object or phenomenon being modeled. Scientists form hypothesis when applying the scientific modelling. When actual experiments are hard or impossible to carry out due to constraints in the real world, such as testing different ways that human disease spreads, computing models and simulations are often used.

Hypothesis are refined by examining the insights that models and simulations provide into the object or phenomenon. As an hypothesis is explored the use of computing models and simulations may generate new knowledge and new hypotheses related to the phenomenon being modeled.

In addition to allowing experimentation when real-world tests are prohibitive, computing-based models and simulations very often allow much faster and much less expensive experimentation. The time required for simulations is impacted by the level of detail and quality of the models used. For instance, modeling the spread of disease by a computer simulation that shows humans as dots randomly moving on the screen and spreading disease if a diseased dot comes within 5 pixels of a an non-diseased dot is a very fast simulation because it has a poor quality model compared a model that reflects actual human behavior (eating, sleeping, using buses, using elevators, etc). Fortunately, because models are made of computer software, rapid and extensive testing of models allows them to be changed to better reflect the objects and phenomenon being modeled by simply re-programming the software.