Knowledge is profit!

Simulation modeling platform

We're developin applied simulation models on our own applications program, using semantic technologies for model management.

Simulation modeling is suitable for solving optimization tasks which cannot be solved analytically, such as:

  • Logistic tasks - transportation management, passengers flow management, service management,
  • Technological schemas and processes optimization,
  • Optimal investment, construction and other programs development.
The common task of all optimization models is the search for solutions which assures the best result with minimal costs.

Building simulation models managed by ontologies has the next advantages:

  • Unrivaled possibiities of controlling objects behavior and interactions, not limited by any paradigm.
  • Any physical models may be composed, from the electronic schemas to the hydrodynamics, from heat technics to the quantum physics. Social, economical and any other model types can be developed and interconncted between each other.
  • It is possible to introduce qualitative parameters into the model, including those which are describing subjects behavior - for example, their goals and value orientations.
  • Automatic inference functionality may be used for logical computations during model execution, generating the new knowledge and even change the model itself in the run time.

The simulation modeling platform has the following principles:

  • An ontology is created to define classes of agents and static elements of the modelled system, their properties and relationships.
  • Behavior of each agent is described using a special script language, which offers widest range of the agent behavior control using logical and mathematical calculations.
  • Source data can be loaded into the model using integration tools or entered manually.
  • The simulation is run with defined range of variable conditions. During the simulation, the combination of conditions has to be found, which produces the result that optimally fits the defined target criteria.
Model calculation may be performed in parallel mode on several nodes of computational cluster. If run on the single server, the load is distributed between processor cores.