Integrated Qualitative and Quantitative Decision Support System for Risk Assessment
Background
Risk assessment is a complicated subject where risk is determined by numerous factors, including human error. The general area of risk assessment is vast, with many methods and tools available to use for assessing risk of various environments [1, 2]. Many risk assessment techniques currently used in risk management are comparatively mature tools. However, there are still some limitations for these techniques to be widely and effectively applied to provide useful solutions for risk-based decision making. This is due to the following problems among others:
(1) Under prevailing circumstances, only limited or no previous experience exists on a hazard, for which the statistical accuracy is poor, in particular for a risk scenario where undesired events are extremely rare.
(2) It is extremely difficult to quantify the effects and consequences of hazards as they involve too many factors with a high level of uncertainty, even in those cases where the physical processes are clearly understood.
(3) It is extremely difficult to generate a mathematical model to represent and describe the risk behavior of disaster hazards as risk is a multiple-level and multiple-variable optimization problem.
(4) A large number of assumptions, judgments and opinions are involved subjectively in a risk quantification process. Therefore, it may require considerable skill for a risk analyst to interpret the results produced.
Research Program
This project aims to build an intelligent decision-support system for risk assessment, capable of handling a broad range of risk scenarios via processing non-uniform data. The project is focused on: 1) developing a new knowledge representation architecture suitable for handling heterogeneous information, which is able to organize and integrate the statistical objects (information from measurements, mainly quantitative in nature) with domain knowledge (processing perception, mainly qualitative in nature) into a coherent whole; 2) exploring and deploying state of art data mining methods for generating the decision models from data and domain knowledge; 3) developing new information fusion algorithms based on the new knowledge representation architecture with the aim of making a decision, i.e. risk prediction and classification in risk scenarios; 4) providing a rational analysis and interpretation via the visualized interactive framework where the above three modules will be integrated to assess hypotheses according to the current context as soon as new information is available.
Anticipated Outcomes
This research will build on the existing research work of the supervisors. It is anticipated that a successful completion of the project will result in an integrated qualitative and quantitative decision support system with wide-ranging applicability in safety or security related risk assessment or risk assessment in health care.
References
[1] J. Liu, L. Martinez, H. Wang, RM. Rodriquez, and V. Novozhilov (2010), Computing with words in risk assessment. International Journal of Computational Intelligence Systems, 3(2010), pp. 396-419.
[2] P.A. Ralston, J.H. Graham, and S.C. Patel (2006), Literature Review of Security and Risk Assessment of SCADA and DCS Systems, T.R. TR-ISRL-06-01, Editor. Intelligent Systems Research Laboratory, Dept. of Computer Engineering and Computer Science, University of Louisville: Louisville, KY 40292. 2006, pp. 20.
Personnel Involved
First Supervisor: Liu, J Dr
Second Supervisor: Glass, DH Dr
Collaborator: Prof Luis Martinez-Lopez (Spain)
Collaboration: This project does not involve collaboration with another establishment
Synopsis:
Risk assessment is a complicated but important subject with many methods and tools available to use under various environments. However, there are still some limitations for these techniques to be widely and effectively applied to provide useful solutions for risk-based decision making. This project aims to build an intelligent decision-support system for risk assessment, which is capable of handling a broad range of risk scenarios and, and able to organize and integrate the statistical objects (information from measurements, mainly quantitative in nature) with domain knowledge (processing perception, mainly qualitative in nature) into a coherent whole. It is anticipated that a successful completion of the project will result in an integrated qualitative and quantitative decision support system with wide-ranging applicability in safety or security related risk assessment or risk assessment in health care.