Sometimes you need to perform experiments with the system in order to understand its behavior, test and compare different scenarios, or find optimal solutions. However, there are systems with which you cannot experiment in the real world because it would be too expensive or sometimes even impossible. In such cases people move from the real world to virtual world of models, perform experiments with the model of the system in a risk-free environment, and map the solution back to the real world.

If the system is significantly dynamic, i.e. if its state changes over time, it has causal and time dependencies, time-related constraints, etc., and is complex (so that it cannot be represented by analytical calculations, by formulas), the only way to explore the system behavior is to simulate its model – build a trajectory of the system in time. The model in this case is a set of rules telling how to obtain the next state of the system from the current state. Depending on the modeling method, these rules can be e.g. differential equations, state diagrams (statecharts), process workflows, etc.

If we consider [dynamic] simulation modeling for business applications, we will find three major methodologies of model development: System Dynamics (SD), Process-centric (“Discrete Event”, DE) modeling, and Agent Based modeling (AB). While the first two were suggested in 1950s and 1960s, agent based modeling has been adopted by simulation practitioners after year 2000, but since then has accumulated a good number of success stories. Both SD and DE modeling employ system-level (top-down) view on things while AB approach is a bottom-up one: here the modeler focuses on behavior of the individual objects.

The system dynamics method assumes high abstraction level and is primarily used for strategic level problems. Process-centric (“DE”) modeling is mainly used on operational and tactical levels. Agent based models are used at all levels: agents can be competing companies, consumers, projects, ideas, or vehicles, pedestrians, robots, etc. Unfortunately, the people who develop SD models (say, of the market dynamics) rarely talk to people who use DE or AB (e.g. of production or supply chain) – they simply talk different languages. What further separates those communities is the tools they use: all traditional tools are designed to support either SD, DE or AB modeling paradigms.

The main idea behind AnyLogic is to bring those people together by putting all three modeling methods on the same platform. There are two major benefits of using AnyLogic:

• You can easily vary the level of abstraction and viewpoint until it perfectly fits the problem. With AnyLogic you can finally say “No” to workarounds. You do not need to fight the modeling language and the tool anymore.
  • If you feel the SD abstraction level (stocks, flows, and feedback dynamics) is enough for the problem, use aggregated view.
  • If the system can be naturally represented as a process (sequence of operations, entities, resources) – use DE.
  • If you are more comfortable with specifying individual behavior of objects (people, vehicles, companies, assets, projects, etc.) – use AB modeling.
• And, you can mix different methods in one model

Making AnyLogic your company-wide simulation platform will enable you to gain deeper insight into complex interdependent processes going on inside and around your organization.