Management Science and Operational Research Blog

Systems Dynamics (SD) is one of the lesser known disciplines in Operational Research. It has been invented by Jay Forrester¹ at the MIT and aims at the analysis of systems whose complexity lies within its feedback structure. System Dynamics has been proven to be a useful tool if it comes to strategy and policy evaluation.

The System Dynamics working hypothesis assumes that the behaviour of a system with feedback is difficult to predict. So a computer simulation of stocks and flows supports the analysis of external influences on the system. The Bullwhip Effect is one famous example of such effects², even though the analysis is based on mathematics rather than simulation in this particular case.

Collaborative Systems Dynamics

Most SD projects establish a classical ‘client-doctor’ relationship³. The consultant interviews key personnel, retreats for some back-room work and then reappears with a well elaborated computer simulation. So the question arises, whether it is possible to democratize the model building process. Would it be possible that each work-group within a company builds its own model? Is there a way to connect these individual models to one grand simulation of the companies activities?

I belive that the dissemination of client-server technology enables the development of such a collaborative system quite easily. For example, each workgroup could build a model of their activities on a server. These sub-sytems could then be linked to one unified model, which reveals the internal dynamics of the entire organisation. An schematic illustration of such collaborative approach is given in the picture below.

This picture makes it clear that certain prerequisites must be given in order to develop an effective Collaborative Systems Dynamics model:

• Employees must be sufficiently trained in the concepts of Systems Dynamics and should have sufficient autonomy in regard to their workplace.
• The existence of clear client-supplier relationships between workgroups is paramount.
• Each workgroup has to have well defined duties within the company.
• The work must be repetitive, at least to a certain degree.
• Due to the nature of Systems Dynamics simulation, the work and information units should be infinitesimal small. This won’t be the case in most circumstances, but the company must seek an continuous reduction of batch and lot sizes.

Such conditions can’t be found in all companies. However, if an organisation wishes to adopt lean management, it has to observe these principles anyhow⁴. So a Collaborative Systems Dynamics project might actually be supportive to the implementation of lean management. Besides that it generates a couple other advantages:

• Unknown dynamics within the company can be unveiled.
• The robustness and responsiveness of the company’s structure can be tested and improved.
• The employees obtain a better understanding of their work and its influence to the company’s output, so that their workplace make more sense to them. This boosts motivation and supports the problem solving process, according to the theory of Cognitive Psychology.
• Managers understand better the influence of organisational structure on the company’s performance and dynamics. The Collaborative Model also allows them to test any modification before putting them to practice.

Since every coin has two sides a couple of downsides and likely pitfalls should also be observed:

• The eagerness of seeing a proper model working might lead to sloppy model verification and validation. However, the collaborative approach might reduce such risks, because any open ends (outputs without inputs between workgroups, or inputs without outputs) strongly indicate inconsistencies and force the modellers to verify their assumptions.
• The employees must be trained and the modelling process usually takes time. Since the entire company is involved in the process the costs of such a collaborative model can be significant.
• The model might serve as the only guide to test hypothesises and implement changes. Systems Dynamics should support the problem solving process but not replace critical thinking and common sense, as it the case with any other Operational Research technique.

In my opinion the prospects for Collaborative Systems Dynamics are looking promising. So the question arises whether traditional Systems Dynamics software, like iThink, can be adapted for the creation collaborative models, or if a new kind of software has to be developed for such a purpose.

1. Forrester, J.W., Industrial dynamics. Waltham, MA: Pegasus Communications, 1961. [↩]
2. Chen, Y. F., Z. Drezner, J. K. Ryan and D. Simchi-Levi, Quantifying the Bullwhip Effect in a Simple Supply Chain: The Impact of Forecasting, Lead Times and Information. Management Science, 46, pp. 436–443, 2000. [↩]
3. Schein, E. H., Process Consultation Revisited: Building the Helping Relationship, Prentice Hall, 1998. [↩]
4. Spear, S. and Bowen, H. K., Decoding the DNA of the Toyota Production System.,Harvard Business Review, September-October, 1999. [↩]