Why Use Multimethod Modeling?

AnyLogic multimethod or multi-paradigm simulation gives you the opportunity to choose the most efficient method or a combination, in order to best address your problem. Whether you are familiar with one or more modeling paradigms, or new to modeling and simulation, employing multimethod modeling capabilities will enhance business process optimization, strategic planning, forecasting and beyond.

Let us first take a look into each paradigm and where they are best utilized, then explain where multimethod modeling fits and how it can improve your modeling process and simulation results.



Various business optimization problems are typically solved by using discrete event modeling. For instance, simulation of enterprise processes, such as production facilities, warehouses, and retail stores, due to their process-centric properties, best described as a sequence of separate events.

Agent based modeling is best if your project or problem surrounds uniting a network, such as a supply chain. In this case, you can model how each component makes decisions and how those decisions impact the other components. Examples of these relationships include knowing how much raw material to purchase due to customer demand, or the number of goods to produce and ship, in which each aspect is modeled as an independent agent. These become multimethod models when the process that happens "inside" the agent is modeled using discrete event or system dynamics.

System Dynamics is used if you want to better understand the correlation between demand fluctuation and manufacturing facility capacity and allows you to model the market as a whole system, where demand is influenced by various factors, like advertising, word of mouth, and other external circumstances. If you want to model the market as a community of people, where everybody makes decisions independently, it is appropriate to use agent based modeling such as this agent-based consumer market simulation. Whether you choose system dynamics or agent-based modeling for your market analysis, you can expand your research by using the supply chain model I mentioned earlier as input data into your consumer market model.

Modeling such complex systems using only one simulation method is a difficult task. With multimethod modeling, the model extension limit depends only on your business needs and, of course, the modeler's proficiencies. Using multiple modeling paradigms enables you to simulate virtually any process and combine multiple models. Furthermore, multimethod makes model building simpler and provides increased flexibility.



For additional information, Andrei Borshchev, founder of the AnyLogic Company, gave on overview of multimethod modeling in this video from a workshop at Winter Simulation Conference 2013.