Reliability, risk and uncertainty modelling

Cranfield University has long standing expertise in the fields of reliability & uncertainty modelling. In late 2009, the Risk Centre acquired a substantial pool of this expertise, adding a vibrant and well-established research portfolio to the Centre’s work.

Dr Shaomin Wu is a Lecturer in risk and decision analysis and leads a team of post-doctoral research staff and three PhD students to further research in the fields of: 

  • Reliability modelling
  • Risk analysis
  • Uncertainty modelling
  • Maintenance/ warranty policy optimisation

Predictive techniques have historically been used to measure uncertainty and reliability in industry; for example in defining warranty policy or component reliability. Shaomin currently has an active research portfolio with industry clients (including Atkins Boreas Consultants Ltd, Yorkshire Water Ltd, and 3Com Corporation), EPSRC and leads a number of novel research contributions to this field.  

A key area of research within Shoamin's portfolio is in "managing risk and uncertainty in warranty servicing policy". This research is conducted by post-doctoral researcher; Dr Artur Akbarov, supervised by Dr Shaomin Wu, and funded by EPSRC:

 

Managing risk and uncertainty in warranty servicing policy

This research specifically addresses the fact that rapid technological advances in many industries, especially in the electronics manufacturing industry, make products obsolete quicker than before. This requires novel approaches to managing risk and uncertainty in making warranty servicing policy. Existing approaches are developed based on estimated warranty claims of each single product but do not sufficiently employ information on the relationships between products. As a result, the global optimisation might not be found and hence huge resources (i.e. capital and raw materials) are wasted. When products are manufactured by the same company and there is competition for fixed budgets, restricted stockroom space or limited maintenance manpower; information on the similarities and competing factors can be very useful in searching global optimisation solutions and hence should be utilised. A literature review of warranty servicing has provided very little information specifically in this area.

This project introduces novel heuristic and hybrid approaches to optimising warranty servicing policy for a batch of different types of products. Models for predicting warranty claims are developed on the basis of sufficiently using information associated with warranty claims, and then a modern portfolio theory based approach is introduced to model risk-sensitive decision making.

 

Finally, Shaomin brings to the Centre an additional statistical capacity for predictive modelling in the environmental risk arena. This is particularly valuable to aid decision-making in circumstances where little data is available and uncertainty is high. 

For more information relating to current and previous research within the project, please view the links to PhD research and MSc research.