曝光臺 注意防騙 網(wǎng)曝天貓店富美金盛家居專營店坑蒙拐騙欺詐消費(fèi)者

University of Birmingham
Edgbaston, Birrningham B15 2TT
Tel: 021 414 4762 Fax: 021 414 4281
Email: sfl @cs.bham.ac.uk
Summary
Insurance brokers Alexander & Alexander are involved with safety-critical and mission-critical systems on a day to
day basis and, together with the University of Birmingham, wanted to research the most suitable support
technologies for risk management.
The approach adopted took the following steps: Evaluation of risk management as a safety-engineering technique.
Analysis of the problem as two streams (although naturally interlinked): domain representation and uncertainty
model. Select several of the more suitable approaches, e.g. KBS, Networks, and Trees for domain representation or
Bayesian probability, Certainty factors and Fuzzy Set Theory for uncertainty modelling. Review these alternatives
theoretically and test them by implementing small models to arrive at preferred solutions.
Nine Case studies were carried out with the Institute of Risk Management on: Risk Analysis, Financial Business
Risk, The Commercial Risk Environment, Health & Safety Risk, Insurance, Physical Risk, Risk Financing, Liability
Exposures and Corporate Risk Management. Research was also carried out on methods of quantitivative and
qualitative analysis, e.g. Dempster-Shafer Theory, Probability Theory - subjective and objective, Fuzzy Set Theory,
and potential technologies for their implementation, e.g. Pearl’s Bayesian networks, Rules with certainty factors,
Fuzzy Causal Maps, etc.
For a more detailed description see Managing risk in Chapter 6.
References
[1] The True Cost of Risk and Its Impact on Safety, S. F. Langley, Technology and Assessment of Safety-critical
Systems, Ed Redmill and Anderson, 1994, ISBN 3-540-19859-8..
[2] Classifying Representations of Risk: The RATIFI Project, S. F. Langley, Safety-critical Systems: The
Convergence of High Tech and Human Factors, Ed Redmill and Anderson, 1996, ISBN 3-540-76009-1.
Project number: 9310 Approx. Project Cost: £0.3m Start date: Feb 1993 End date: Jan 1996
EPSRC Grant number: j18859
Advances in Safety Critical Systems - Results and Achievements from the DTI/EPSRC R&D Programme. Compiled and edited by Mike Falla.
22
RED Reliably Engineered Decisions
Integral Solutions Ltd,
Lloyd’s Register of Shipping,
Masons Solicitors,
Imperial Cancer Research Fund,
Queen Mary and Westfield College
Contact: Mr D Elsdon
Integral Solutions Ltd
Berk House, Basing View
Basingstoke
Hampshire, RG21 4RG
Tel: 01256 55899 Email: davide@isl.co.uk
Summary
Decision theory based on mathematical probabilities is sound, but difficult to apply to knowledge-rich problems
such as health care. Knowledge-based systems can support knowledge rich decision taking, but existing tools and
notations are not sufficiently rigorous for safety-critical applications. RED set out to unite the strenths of both
approaches to provide rigorously engineered decisions.
Outputs have included:
· a model of knowledge-based decsion taking
· an abstract logic of obligation and time, and a concrete representation language
· a study of the safety constructs implicit in a range of published protocols
· several software prototypes
· a proactive legal methodology.
For a more detailed description see section 14.1, Knowledge-based systems and safety critical decisions, and
section 9.3 The legal lifecycle.
References
[1] Towards Rigorous Knowledge-based Decision Engineering, Tom Khabaza, Expert Systems '93
[2] Expert Systems for Safety Critical Applications: theory, technology and applications, IEE colloquium on
Designing Safety Critical Systems, April 1994.
[3] Knowledge based systems for safety-critical applications: theory, technology and applications, John Fox,
Subrata Das, Peter Hammond & Paul Krause, RED Report ICRF/WP/500/6, + C.
[4] Engineering Safety Into Expert Systems, John Fox, Technology and Assessment of Safety-critical Systems,
Ed Redmill and Anderson, 1994, ISBN 3-540-19859-8.
[5] Formalising safety in decision support systems, Subrata Das, Mike Clarke & John Fox, RED Report
QMW/WP/720/8, + C.
[6] Safety and Decision Support in Oncology, Peter Hammond, Adrian Harris, Subrata Das & Jeremy Wyatt,
Meth. Inf. Medicine 33(4) 1994.
[7] Prevention, detection and amelioration of adverse events in protocol-based decision support, Peter Hammond,
Adrian Harris, Subrata Das & Jeremy Wyatt, Proc. AI in Medicine Symposium, Stanford University, March
1994.
[8] Computer support for protocol based treatment of cancer, Peter Hammond & Marek Sergot, Journal of Logic
Programming, 26(2), Feb 96, 93-111.
　
中國航空網(wǎng) www.k6050.com
航空翻譯 www.aviation.cn
本文鏈接地址：航空資料25(156)