Incorporating Multiple Linguistic Variables in System Dynamics Modeling

We propose a method for incorporating multiple linguistic variables into a system dynamics framework. A simple example is used to illustrate the procedures necessary to define linguistic variables using triangular membership functions within the VENSIM Simulation Environment. We illustrate the operations of linguistic variables through a sales and service model where two linguistic variables i.e., customer’s satisfaction with respect to service, and lead time associated with a product, impact the conversion of potential customers into customers. After having created fuzzy triangular membership functions, we obtain the combined effect of the two linguistic variables using the Max-Min inference procedure. For de-fuzzification we use the notion of the Largest of Maximum to translate the fuzzy representation of the combined effect into a crisp value. Finally, we provide simulation results pertaining to the probability of generating new customers and profits by considering pessimistic, optimistic and intermediate fuzzy rules for our model.