http://www.dirc.org.uk/  
 
   
   
Overview
Research
 

   Themes  
   Results

Sites
People
Publications
Events
Related Projects
   
 

Project Activity 4 - Objectives

We shall explore the role of decision support technology via two application drivers. One will investigate dialogue between regulator and licensee about safety-critical systems, using the nuclear industry as example; the other the interface between management and technical staff in a commercial organisation facing changing regulatory and commercial conditions, using Northern Electric as example. This activity, which addresses the evaluation of dependability, and that on Design for Dependability, which in contrast addresses the achievement of dependability, are closely linked, so there will be extensive interaction between them.

The aim of this project will be to develop methods and tools to aid decision-making in all situations where the dependability of IT systems is an issue. An important requirement will be a better understanding of means for marshalling disparate kinds of evidence to support decision-making in domains such as regulation (e.g. via the use of quantitative dependability cases);

  • means of assessing the 'strength' of evidence;
  • means of assessing the dependence of different evidence strands on one another;
  • means of combining different types of evidence when these are very disparate in nature;
  • means of eliciting expert judgements about dependability of software-based systems, taking account of current psychological knowledge of human error; and
  • means for validating quantitative and qualitative assessments of dependability based on disparate evidence (e.g. BBNs);
  • understanding of notions of diversity in dependability arguments (e.g. multiple argument 'legs'), in particular we shall investigate whether the current mathematical models for (design) diversity apply here;
  • understanding of notions of diversity in software (and other) development processes, in particular we shall investigate ways in which process diversity can be deployed most cost-effectively (to increase version reliability, to increase version diversity); and
  • understanding of human failure processes, both in designing and building complex IT systems, and in assessing their dependability.

 
Page Maintainer: webmaster@dirc.org.uk Credits      Project Members only Last Modified: 11 August, 2005