Developing a model to predict time delay in road construction projects using bayesian networks

• Main Author: Mohammad, Almohammad
• Format: Thesis
• Language: English
• Published: 2020
• Subjects: T Technology (General); TA Engineering (General). Civil engineering (General)
• Online Access: http://umpir.ump.edu.my/id/eprint/35240/1/Developing%20a%20model%20to%20predict%20time%20delay%20in%20road%20construction%20projects%20using%20bayesian%20networks.wm.pdf; http://umpir.ump.edu.my/id/eprint/35240/

Time is one of the three leading indicators by which project success measured. As Malaysia is looking forward to becoming an advanced nation, efficient infrastructure is needed. Therefore, completing these projects on time is very important to achieve this goal. However, a considerable number of construction projects in Malaysia have experienced poor time performance. Time delay is considered to be one of the major problems faced by Malaysian construction projects. Thus, this research is carried out to investigate the causes of delay in construction projects and further identify key risk indicators that have a significant effect on project duration. Bayesian networks (BNs) utilized for time-delay prediction by which project status in terms of time can be examined. Scope of this study focus to federal road projects in Malaysia. A literature review was undertaken covering construction projects in Malaysia and road projects in developing countries which resulted in 67 causes of delay divided into 12 groups. Semistructured interview with three expert panels nominated by Public Work Department (JKR) conducted to evaluate the delay causes. A total of 56 causes were determined as relevant to Malaysian road projects. Data collection was then carried out using a questionnaire survey in which respondents were randomly selected. The targeted population was drawn from construction practitioners involved in road construction representing four entities, namely: owner, contractor, sub-contractor and consultant. A total of 500 copies were distributed and 219 valid responses were received. The data were then analysed using relative importance index (RII) for risk frequency and impact. Risk rating (RR) was further established based on the multiplication of both attributes leading to rank the delay factors from the most to least important. Bayesian networks (BNs) were employed to develop a prediction model of time delay based on significant factors causing the delay. The structure and parameters for the BNs model were defined based on knowledge of road experts who have been also approached to verify and validate the BNs outputs. The results indicated that the most significant factors causing delay in federal road construction projects in Malaysia are: financial difficulties faced by owner/ client, bad weather conditions, delay in payment for completed work by owner, material price fluctuation/ increase, cash flow of contractor is insufficient, equipment failure (breakdown), inadequate contractor’s experience, ineffective scheduling and planning of project by contractor, slow equipment movement and slow decision making. The RR value for top ten delay causes ranges between 13.818 related to financial difficulties faced by owner/ client and 9.993 related to slow decision making. In addition, the validation of this model through expert’s opinion confirms that the BNs model is adequate to represent the timeframe of the road projects and can be used for other construction projects with minor modifications. However, it is recommended to apply more reliable methods to identify prior and conditional probabilities for the model to obtain more reliable outcomes.