System Dynamics Modeling to Improve Your Supply Chain
Profit Point has recently added another tool for analyzing and managing businesses and supply chains – system dynamics modeling. System dynamics focuses on feedback effects in complex systems. That is what distinguishes it from other simulation and modeling techniques. Feedback means that although X influences Y, it is also true that Y influences X even if this influence is mediated by a string of causal relations.
Here is a typical System Dynamics drawing from Wikipedia. In this diagram there is a flow of individuals from a category called potential adopters into a category called adopters. The rate of adoption is controlled by the ‘valve’ called ‘new adopters’. Many simulation modeling approaches would treat this as a ‘one-way street’ and attempt to model the system by controlling the valve with some external data in the hope of reproducing behaviors observed in the real system.
The point of System Dynamics is to explicitly include the feedback which causes the observed behavior. The model above could be a portion of a model of product adoption for the pharmaceutical industry. Another classic example is a predator – prey population model. When prey are plentiful, the predator population increases, which reduces the prey, which in turn reduces the predators. The extension of this classic model to a system of a business and its resources is fairly natural.
The System Dynamics approach has also been widely applied in typical supply chain scenarios. The effects of feedback can be studied in the context of inventory management and shipping policies. The figure below is of a classic System Dynamics approach to modeling the supply chain from raw materials production through shipment of the finished goods. Feedback is represented by the directed arrows from upstream processes back to the downstream processes. In many cases these represent control mechanisms meant to keep the system in balance or within limits. Successful feedback usually dampens oscillatory behavior – stock-outs, long supply, or obsolete inventory items, for instance.
System dynamics models are classified as continuous simulation models and are typically built in languages specifically designed for these types of models. They are quite different from the usual simulation models that are based on discrete entities and systems that jump from event to event in time. System dynamics models are based on equations that describe (usually) continuous flows of material, people, money, influence, etc. The models are highly useful as policy analysis tools – revealing the often unintended consequences of business rules. The models can examine periods of years or decades, while including external effects, e.g. the economy, and all the significant interactions within a business and with the outside world.
We look forward to bringing this exciting, powerful, and well-established analysis tool to bear on problems and opportunities of our clients.