Keywords: risk management, risk identification, qualitative analysis, case-based reasoning, knowledge management, information technology.

Risk management is an important tool used to aid the understanding of the scope and potential problems related to projects to support decision making. Risk identification is the fundamental stage for success risk management as it decides whether or not the further risk analysis is necessary. At present, well known risk identification methods largely rely on the experience of project experts. However, subjective judgement and intuition usually accompany the expert opinion. In addition, sharing and transferring of this knowledge is restricted by the experts' availability and job changing. Case-based reasoning learns from experience, which matches the human techniques that catch and process information and knowledge in relation to project risks. It is hoped that this methodology can improve project risk identification and support further decision making.

Currently, risk identification is largely based on the subjective judgement of humans. The main purpose of risk identification is identifying uncertainty and opportunity, while uncertainty sources from lack of knowledge. Hence, knowledge accumulating and sharing is the key to break the bottleneck of risk identification. Psychological research shows that human make future planning decisions generally based on the three most recent decisions made by the same manager. According to Pender [1], it seems that about nine decisions attributes that a person can effectively encompass each time, which illustrates that managers have a limited information processing capability. Consequently, they cannot directly deal with complex problems even though the information may be available in some form.

Case-based Reasoning is a problem-solving paradigm that solves a new problem by remembering a previous similar situation and by reusing information and knowledge of that situation [2]. Case-based reasoning has the advantage to handle non-numerical information. On account of the complexity of construction project, simple modelling system is difficult to carry out risk identification. Compared with traditional risk management methods, Artificial Neural Networks (ANNs) have the advantage of self-learning, self-organizing and real time operation. However, ANNs are like a black boxes that cannot provide explanations of their decisions. Knowledge-based systems must explain their decisions by referencing to their rules which the user may not fully understand or accept and the rule processing is time consuming. Case-based reasoning is a paradigm to learn from previous experience and solve similar problem by referencing earlier solution. The reasoning process is retrievable and it can deal with textual information easily, so the output solution might be more easy to understand and acceptable.

Information from previous projects is processed and stored in a knowledge base. A new project is first entered into the project description form as the cases stored in the knowledge base. It then searches the knowledge base to find similar previous projects, reuses the risks identified, and revises these risk lists to come up with suggested risks for the project in question. This preliminary suggestion is reviewed by human experts, and then a new final report of identified risk for this project is generated. This solution is associated with the project description and is retained in the knowledge base as a piece of new knowledge for future use. The final process called refine is more about the maintenance of the knowledge base, updating the information, such as when the project finished, so that the identified risks and response strategy can be evaluated and the knowledge base is revised.

The development of a prototype system CBRisk in introduced in detail and the validation and verification scheme is also demonstrated. This paper concludes that the CBR as a paradigm to simulate the process of human solving problems is particular useful coping with complex information and knowledge of risks in relation to construction project. CBRisk is a prototype that implements CBR in modelling risk identification for construction projects. It does not try to take over the role of human experts or force them to accept the output of the system; instead, it aims to provide more relevant evidence to facilitate human experts making final decisions. This approach intends to exploit the advantage of knowledge sharing, increasing confidence and efficiency in investment decisions, and enhancing communication among the project participants. This should bring about a focus change to risk identification and promote the application of risk management to the construction industry.

1

Pender S., "Managing incomplete knowledge: Why risk management is not sufficient", International Journal of Project Management, 19 (2), 79-87, 2001. doi:10.1016/S0263-7863(99)00052-6

2

Aamodt, A., and Plaza, E., "Case-based reasoning: foundational issues, methodological variations, and system approaches", AI communications, 1 (7), 39-52, 1994.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £80 +P&P)