Document Actions

Research Field: Multi Level, Multi Formalism Modelling and Variable Structure Models

last modified 2006-04-03 09:57

Modelling and simulating complex systems whose sub-systems scale over several orders of magnitude with respect to size, numbers, or time requires switching between overview and detail perspectives on need. These multiple perspectives are supported by multi-level and multi-formalism modeling and simulation. Variable structure models support models that change their own composition, interaction, and behavior pattern during simulation. All these approaches lend additional structure to the description of systems.

Multi-level models might imply

Descriptions at different temporal scales: allow to model both slow and fast sub-systems of a dynamic system within one model.
Descriptions at different abstraction levels are supported by the composition of models, and by variables that scale over various orders of magnitude. Compositions allow to structure according to function and to distinguish between individuals. When quantitative variables are transformed into qualitative ones, certain ranges of values are aggregated, introducing categories.

Descriptions at micro and macro level capture the dynamics of interacting individuals as well as that of entire populations. Thus, the macro level forms also an aggregation, i.e. an aggregation of models representing individuals all belonging to the same population.

Discrete event approaches focus typically on a micro perception of systems and their discrete, stochastic dynamics. Variables can be arbitrarily scaled. Sub-systems advance over a continuous time scale, but in steps of variable size. Discrete event-based modeling formalisms appear well equipped to form a basis to address the challenges of multi-level
modeling.

However, as different modeling formalisms even for discrete event systems simulation (to say nothing of the many formalisms for hybrid systems and continuous systems) exist their combination might help in structuring the knowledge about a given system and in efficiently executing the model.
Multiformalism modeling and simulation is either supported by

model transformation
model integration by specialized simulation engines
simulator interoperation

Variable structure models which asses and access themselves, lend additional structure to the temporal dimension of models. Phases of similar structure are distinguished supporting the evolution systems with varying behavior, interaction, and composition pattern.

Related Research Projects:

2002-09-30 - 2005-03-31	DELMS - Discrete event life science modelling and simulation
Project coordination: Adelinde M. Uhrmacher

Sections

Personal tools

Document Actions

Research Field: Multi Level, Multi Formalism Modelling and Variable Structure Models

Related Research Projects: