ESD.83 Knowledge Domains in Engineering SystemsSystem Dynamics: Tackling the World’s ComplexityRandy UrbanceIntroductionSystem Dynamics was a field that was created to handle the complexity of realproblems. In an interconnected and changing modern world dealing with complexsystems can be a daunting task. System Dynamics’ basic premise is that therelationships between factors in systems exhibit non-linear behavior. These non-linearrelationships are often compounded by several dynamic mechanisms includingcomplexity and information delay. As a result the behavior of a complex dynamicsystem is inherently difficult to understand and foresee. System Dynamics was amethodology created to cut through the complexity of systems, establish manymathematical relationships between system components and organize the relationshipsinto a whole system model. Simple relationships are built up and added together to geta better picture of how a system behaves. This field of study, while originating as amethod to tackle complex world of business management, has found applications inevery aspect of life from government policymaking, human biology, and globalenvironmental studies. Many academics, policy makers and business leaders havebeen won over by the power of System Dynamics as a tool for understanding the world.This short paper describes the history, basis and attractiveness of System Dynamics asa tool for engineering and re-engineering complex systems.HistoryThe field of System Dynamics (or SD) was founded in the 1960s at theMassachusetts Institute of Technology by Prof. Jay Forrester. Prof. Forrester hadgained wide respect for his work in the field of digital computing following World War II.His development of the Wirlwind computer, the backbone of the SAGE system, was avital contribution to the United States air defense. Despite his accolades Prof. Forresterbecame restless with his position in the electrical engineering department. Whenoffered and opportunity to switch to MIT’s Sloan School of Management, Prof. Forresteragreed enthusiastically. The Sloan School’s affiliation with MIT’s technical roots hadprompted them to look for more ‘scientific’ method for business management.Forrester’s views on electrical control theory and feedback fit well with the controlling ofbusiness. Prof. Forrester translated his knowledge of electrical flows and digitalcontrollers in to a generic mathematical language that could easily be translated tocapital flows and inventory control. Forrester summarized his methodology in BusinessDynamics published in 1961. This text modeled and explained many modes ofbehavior viewed in business systems, and was well received in business and academia.Following the release of Business Dynamics SD gained more supporters and Prof.Forrester created a whole department at MIT devoted to the method and itsapplications. This group lead the way for further SD research into other systems likeUrban Dynamics, that describes and models social, labor and population dynamicsobserved in cities, and World Dynamics, delving into the ambitious pursuit of modelinghumanity and its effects on the planet.Basis for SDThere are two major behavior modes of systems explored by System DynamicsModels:1) Positive Feedback (exponential growth or decay): Many dynamic systems donot interact in a linear manner. In a dynamic model, positive feedback reinforcesgrowth (or decay) with a looping structure. A population of chickens (or anypopulation) is a good example. Chicken lay eggs, which hatch into morechickens, that lay more eggs. The chicken population will grow very rapidly in anexponential manner. Many systems exhibit this behavior mode from biologicalpopulations to bank interest.2) Negative Feedback (balancing or equilibrium dynamic): Other dynamic systemstry to maintain or balance a system. A good example of this dynamic is abiological population near its carrying capacity. For example, imagine apopulation of deer is close exhausting a forest’s supply of food. Any additionaldeer will limit food to the others and more will die. If deer migrate away from theforest, more food will be available for the current population and more offspringwill be able to survive, increasing the population. This negative feedbackenforces and maintains an equilibrium state.Deer PopulationAvailability ofGreen FoodPopulation ofChickensNumber of Eggs++-++-Additional Insights from SDUnderstanding the important feed back structure was the basis for much ofForrester’s first studies of dynamic systems. However, there are several aspects of realsystems that had to be tackled before it was a robust method. When interacting withcomplex systems, humans normally have to observe or measure the current state of thesystem. A problem that arises is that no measurement or observation is instantaneous.Data must be collected, analyzed and action plans must be formulated. Wheninteracting with dynamically growing systems this measurement delay can bedangerous. If one is trying to control a rapidly exploding population, by the time yourealize that the growth is too much it may be too late. This usually resulted in a boom-bust dynamic that is widely observed in real systems. Modeling delays also allowmodelers to understand how seemingly simple systems will have oscillatory behavior.Complexity also has a major impact on the working of a System Dynamicsmodel. In the final formulation of a large SD model many different feedback loops willbe linked together. While each single loop relationship is inherently simple to describethe overall behavior mode of the system may be unforeseeable. The act ofdecomposing and recomposing the relationships of complexity is another strength of theSD methodology.Applications of System DynamicsNon-linear growth, information and observation delay and system complexity arevery difficult for the human mind to envision. With Jay Forrester’s original work in digitalcomputing many of these complex issues could be address automatically usingcomputers. Combining his two fields of expertise Forrester embarked on a series ofmodeling studies in several important fields.In his first SD text, Business Dynamics, Forrester outlines how natural businesssystems for handling product orders, production and inventory can show oscillatorybehavior, even in the presence of stable demand. Forrester also showed how orderingand production scheduling decisions coupled with