5.8. Appendix 1: Model Communication and Policy Structure Diagrams

In this chapter I have used the factory model to illustrate the iterative steps in the model building process from problem articulation to simulation. It is a small model containing 22 concepts in total (two asset stocks, four flow rates, six converters, one built-in smoothing function, eight constants and one exogenous time series input) that just about fits on a single page. Nevertheless, as Figures 5.8, 5.17 and 5.18 show, there is considerable visual complexity in the model, which can inhibit communication. Communication is even more difficult with practical application models that contain hundreds or even thousands of concepts. There is a need for visual clarity, which can be achieved by grouping model concepts into meaningful clusters.

The factory model was divided into two main sectors, one for production control and the other for workforce management. Within the two sectors are the main operating policies that collectively guide and coordinate the accumulation of asset stocks. These policy functions can be indicated on a stock and flow diagram by using shading, as illustrated in Figures 5.8 and 5.17. Alternatively, much-simplified pictures can be drawn that focus on operating policies and strip away all the underlying formulation detail. The result is a 'policy structure diagram'. Figure 5.34 is an example. It shows the factory model in terms of its main operating policies and stock accumulations. Here, an operating policy is represented as a large circle with information inputs and outputs. In some cases, the policy sits alone in a sea of information, a decision-making process – like the demand forecast or inventory control – that interprets information and hands it on to other parts of the organisation. In other cases, the policy subsumes a flow rate where a decision-making process, such as hiring, leads to action (the hiring rate).

Figure 5.34. Policy structure of the factory model

Policy functions are similar to converters except that the transformations they bring about are more complex than can be expressed in a single conversion or a single algebraic equation. Policies involve several judgemental and computational steps. For example, inventory control involves goal formation (desired inventory), sensing a gap (the difference between desired inventory and actual) and corrective action (the amount of additional production to schedule in order to close the gap). Hence, the large circular symbol with horizontal lines represents a complex, multi-step conversion process.

The policy structure diagram provides a useful overview of factory coordination without the full formulation detail. In fact, a wide variety of alternative formulations, with different steps of computation and different intermediate concepts and constants, might equally well fit within the framework of the diagram. This flexibility makes it easy to discuss the diagram with people who know factory operations. Twenty-two concepts are reduced to only 11. Moreover, the vital information feedback network is formed from only six policy functions, each with a practical meaning to factory people. I have found such diagrams very useful in my own professional work, both for model conceptualisation and communication. I will continue to use them throughout the book.