Since Vietnam implemented its policy of reform and opening to the outside world, and with the continuous development of its national economy, the Vietnamese building enterprises have been growing steadily in quantity, and the building industry has been increasing its total output value. With Vietnam going to entry into the World Trade Organization (WTO), the process of opening the Vietnamese construction market to the outside world will be accelerated. Moreover, because of economic development, infrastructure has to be improved to adapt with the changing. Many office buildings, apartments, bridges, roads… have been built. However, Vietnam is still undergoing reform of its system, from a planned economy to a market economy. Laws, regulations, market and government management models and systems are still not stable. National market is still controlled by government. All are obstacles preventing contractors from fully understanding the entire situation and risk in the Vietnamese construction market, especially international contractors. Therefore, identify and analyze risk factors in Vietnam construction market is necessary.
Construction projects involve hundreds or even thousands of interacting activities, each with cost, time, quality, and sequencing problem. The costs and durations are uncertain and one response, still surprisingly common, is to shy away from uncertainty and hope for the best. Another is to apply expert judgment, experience, and gut feel to the problem (Roger Flanagan & George Norman, 1993). Construction projects are unique arenas in which highly complex, uncertain and creative projects have to be realized (Hartman, 1998). However, construction project risks are often ignored by most owners and contractors. As a result, unnecessary long and disruptive delays turn an otherwise profitable project into a financially ruinous undertaking (WONG, 2006).
Obviously, there are some overlaps between these risks. Edwards and Bowen (2005) used a source system-based approach to classify risks under two primary categories: natural systems and human systems. The sub-category of natural risks includes events originating in weather, geological, biological, physiological, ecological and extraterrestrial systems. The sub-categories of human risks comprise social, political, cultural, health, legal, economic, financial, technical and managerial systems. Edwards (1999) used the same risk sources as a primary means of categorization to minimize confusion.
Conventional risk analysis techniques, such as Monte Carlo analysis, provide tools to help practitioners to assess impacts of uncertainties, to support the determination or the assessment of the risk level of a project, and to allocate a contingency associated with the possibility of success. Unfortunately, the effectiveness of using this technique is heavily dependent upon experts' opinions and judgments (Xiaoying Liu,1998).
Neuron-Fuzzy is an approach that is free of mathematical models. It requires less expert opinion and judgments than do other techniques. It represents an attempt to simulate the human brain's learning process through massive training. It is able to learn from samples. Knowledge learned is stored within the network. This technology provides a powerful and robust means to assess uncertainty through learning and capturing general patterns in available data. NF integrates both neural networks and fuzzy inference systems. These model frameworks possess both the learning capability of neural networks and the structured knowledge representation employed in fuzzy inference systems (Jyh-Shing Roger Jang, 1992).
On time and within budget are two main outputs of successful project. However, there are risk factors affecting the construction projects’ time and cost. In order to achieve two major project outputs, management team needs to access and analyze those risk factors as a proactive plan. NF is considered as the advantage method to help practitioners who are either not much experience or expert develop a proactive plan to modify threats which possible impact on project performance.
Mr. Pham Hiep Luc made a research on “Application Of Neuro-Fuzzy Networks To Forecast Cost And Duration Variance For Building Projects In Vietnam” which major objectives were to: (1) identify main risk factors which affect the duration and cost variance of building projects in Vietnam; (2) develop a Neuron-Fuzzy model to predict project cost and time variances; and (3) compare Neuron-Fuzzy model with conventional method.
Significant risk factors impacting cost and time variances of building projects in Vietnam were identified in this study. Manager experience, construction method, type of clients, client changes, project complexity and market price fluctuation were identified by respondents as having high correlations with time and cost variances. Factors related with client were given high marks by respondents. However, when choosing the factors which had high correlations with time and cost variance by forward regression technique, the results was quite different from the respondent’s perception. Client type, project priority, vagueness in scope and construction method had strong correlation with time variance. Project function, location, contract type, project complexity, market price fluctuation and project priority had strong correlation with cost variance.
The goal of this result was to apply the Neuro-Fuzzy to risk analysis. Most traditional techniques are heavily dependent upon expert judgment and experience. In complex situations, these techniques are difficult to use because no mathematical model can be applied and the correlations between risk factors are not easy to identify. The weaknesses require the researchers to find out a new technique to approach better the complicated situations in project time and cost risks.
An attempt has been made to apply neuro-fuzzy network in the assessment of risks at the early stage of a project. The relationships among risks, project characteristics, decisions, and outcomes were captured by neuron fuzzy networks. Intelligent models were then developed and tested. They can be used to predict project cost and time variations.
The research results showed that the neuro-fuzzy networks outperform conventional techniques such as multiple linear regression analysis. The practical application of neuro-fuzzy network technology in project risk analysis is promising, especially in the front end stage.
Neuro-fuzzy network (NFN) model is able to capture the risk patterns of projects by learning from historical project samples and to generate a reasonable prediction of project cost and time variances. It is superior to conventional techniques as multi-regression technique and neuron network technique.
Neuro-fuzzy network model with forward stepwise regression analysis provides more accurate estimates of project cost and time variations than NFN with variables ranking by respondents. It also significantly reduces the training time and increases the training efficiency of the networks.
Neuro-fuzzy network model is superior to neural network because, neural network is difficult to determinate its configuration.
This research built a rational base for developing a decision support system to assist project managers (decision makers) in better decision making.
Overall, projects will have a greater chance of success, in terms of “within budget” and “on time”, when project managers direct more effort into managing the identified important factors during project planning at the front end phases and NFN is the useful technique to support decision making process.
His thesis abstract is copied and posted.
The importance of decision making in time and cost estimation for investment processes points to a need for an estimation tool for owners, designers and project managers. This research is aimed to explore the applications of neuron-fuzzy network technology in project risk analysis and to develop models to predict cost and time variances at the front end stage of building projects.
Seventy finished building projects in Vietnam replied by the respondents were used for training and testing the model. Important risk factors at the front end stage are determined by respondents’ ranking. To develop the neuro-fuzzy network model, forward stepwise selection technique was used to identify input sets. Different number of data was being used to train the model. The results of neuro-fuzzy network models were superior to conventional technique (such as multiple linear regression) and neural network models in the prediction of project cost and time variations. Moreover, the result showed that using forward stepwise technique to determine the input set for the model is better than respondent’s perception. One proposed program written by Visual Basic Macro language was conducted in order try to interpret the results with practitioners.