# CPM toolbox

**Background**

Technical products are composed of many interconnected parts that work together to realise the products’ functions. Change made to one part of a product tends to affect other parts and propagate. For instance, a change made to the geometry of one component may require the geometries of adjacent components to also be changed, so they will still fit together. In turn, these changes may require yet more components to be changed. The propagation of change between related elements can also be observed in other types of system, such as processes, organisations, etc.

**The Change Prediction Method (CPM)**

CPM is a DSM-based numerical approach for predicting and analysing how changes are likely to propagate through a system. The design of the system is decomposed into its components and the dependencies between them are captured and quantified. As a result, a product design is modelled as a DSM (and a corresponding numeric network), where the DSM-elements (nodes) represent the components or subsystems and the DSM-links (edges) their dependencies. Likelihood and impact values inside the DSM-cells (on the edges) represent the strength of the linkages with regard to propagation of engineering changes. Afterwards, a stochastic algorithm is applied to calculate the combined risk of change propagation between components considering multiple steps of direct and indirect change propagation. Based on this product model and the combined risk values, CPM generates different diagrams (i.e. Change Propagation Tree, In/Out Risk Portfolio, etc.) to support change propagation analysis.

**Using the toolbox**

**Open or create a CPM model:** Use the CPM-icons at the bottom left side of the screen to insert DSM-elements and quantify the dependencies between them or import a model from MS-Excel.

(click screenshots to enlarge )

**Calculate combined risk values:** Click on CPM-icon to apply the CPM-algorithm and generate a combined risk matrix. The change path length is set to six steps by default and can be modified. By default, CPM produces a general risk profile considering single changes – one at a time – to each component. However, if multiple components change simultaneously, they can be selected accordingly to produce a specific change profile. As result, a new window opens including the combined risk matrix and additional icons for diagrams.

**Analyse change propagation:** Use the diagram icons to generate the following diagrams.

**Case Risk Plot** depicts for a selected change instigating (or receiving) component all affected (or causative) components in a square diagram with combined impact on its vertical axis and combined likelihood on its horizontal axis. The diagram axis can be logarithmic scaled using the corresponding column (or row) of the combined risk matrix (Local Ln-Scaled) or using the whole combined risk matrix (Global Ln-Scaled).

**Change Propagation Path Diagram** depicts for a selected change initiating component all possible change propagation paths of a selected length to a selected target component (TC).

**Distance Network** depicts for a selected change initiating component (represented in the centre of the inner circle) the shortest path length to each affected target component (represented on the circle borders of different order).

**In/Out Risk Portfolio Plot** depicts all components in a square diagram with incoming risk (row average of the combined risk matrix) on its vertical axis and outgoing risk (column average of the combined risk matrix) on its horizontal axis. This allows differentiating between change propagation absorbers and multipliers.

**Product Variant Portfolio Plot** depicts all components in a square diagram with incoming impact (row average of the combined impact matrix) on its vertical axis and incoming likelihood (row average of the combined likelihood matrix) on its horizontal axis. This allows differentiating between types of variants.

**Risk Network** depicts for a selected change initiating component (represented in the centre of the inner circle) the imposed combined risk values of a selected path length to each affected target component (represented within the outer circles of decreasing values). The distance of the components to the central component corresponds to their reciprocal combined risk values.

**See also**

Simulation to evaluate the benefits of change prediction tools

**More information**

CLARKSON, P.J., SIMONS, C.S. and ECKERT, C.M. 'Predicting change propagation in complex design' *ASME Journal of Mechanical Design,* 126 (5), 788-797, 2004.

SIMONS, C.S. 'Change propagation in product design', M.Phil Thesis, Department of Engineering, University of Cambridge, 2000.

The CPM toolbox was implemented by Nicholas Caldwell.