Change Prediction Software Tool

Research Theme: Change Management

Change Prediction Method (CPM) is a tool for identifying relationships among entities (which may be physical components, team members, tasks, etc.) and determining the criticality of the interconnections. CPM can be used to create models (in the form of matrices, node-link diagrams, or component lists), record information about impact types, likelihoods, etc., and display change prediction results and component connectivity using multiple visualisation modes.

Motivation

Most designs are modifications of existing products. During tendering designers and managers also need to quickly assess the implications of changes. For complex product, there is a clear industrial need for a software solution that supports the prediction and change management.

Objectives

  • Give designers and managers the ability to assess the risk associated with a change before it is implemented
  • Visualise the possible effects of changes communicated to other stakeholder
  • Provide an overview over the product architecture
  • Enable designers to explore the architecture of the product

Method

The tool is developed in close collaboration with industry in rapid feedback cycles. CPM tool is implemented as a Java application which can load and save files in XML format. An HCI methodology is followed in order to find and test human-computer interfaces that benefit the design process.

Findings

New visualisation algorithms as well as improved change prediction algorithms have been implemented and tested throughout the development of the software tool. The use of multiple views allows the exploration different "what if" scenarios to identify means to reduce risks resulting from unexpected change propagation.

Acknowledgements

Support and feedback is provided by:

Support for this project is provided by EPSRC.

Selected Publications

  • KELLER, R., ECKERT, C.M. and CLARKSON, P.J. (2006) 'Building connectivity models in design: representations and tools to support cognitive preferences' in Design Computation and Cognition 06, Eindhoven
  • KELLER, R., EGER, T., ECKERT, C.M. and CLARKSON, P.J. (2006) 'Matrices or node-link diagrams: which visual representation is better for visualising connectivity models?' in Information Visualisation, 5 (1), 62-76
  • ECKERT, C.M., KELLER, R., EARL, C.F. and CLARKSON, P.J. (2006) 'Supporting change processes in design: complexity, prediction and reliability' in Reliability Engineering and Safety Systems, (accepted)