The Study On System-dynamic-based Bayesian Networks For Construction Schedule Risk

Construction projects are comprehensive social activities, with large investment,long time, high technical requirements and complex system, there are many uncertainfactors in each aspect. Effective risk management can reduce uncertainties, ensure theproject implemented smoothly. The core of risk management is risk identification,risk assessment and analysis. Only taking the correct recognition and evaluation,appropriate means can be select to prevent and control risks. Many scholars havemade many studies to identify construction project risks, the common methodsinclude the Delphi method, brainstorming method, expert interview method, risk list,etc.. However these methods are usually lack of systematicness and pertinence, riskfactors identified by these methods are independent with each other, which is notaccording with the engineering practice. In addition, risk management is an importantpart of enterprise and project management, its costs attribute determines that riskmanagement must seek utility maximization within limited resources. It is necessaryto understand the action mechanism of risk factors, thus risk management can be moretargeted.Some researchers have discussed the mutual influence between risk factors andeffects by building simulation models. They find that risk factors are not completelyindependent, and their effects on the project cannot be added simply. But theseresearches are just at the beginning, deeply studies must be taken in the future.Schedule risk is one of the most common risk in construction projects, relatesclosely to the project quality and cost, takes a very important role to the project. Thispaper will identify and analysis the factors which can influence the project schedulesystematically, explore the influence mechanism these factors made to the schedulerisk, then build system-dynamic-based Bayesian networks to reflect risk events andthe affected activities and measure the risk probability accurately, so that can providesupports to risk management decision. This study will also develop the theory of riskmanagement, has strong theoretical value and practical significance.During the research process, literature analysis has been taken firstly to identifythe related factors affecting the project schedule risk. On the basis, reason treeanalysis and result tree analysis has been taken to identify the factors further systematically. According to the causal relationship between risk factors, a systemdynamics model of project schedule risk has been built. One or more acyclic graphcan be extracted from the SD model based on the project environment, thenprobabilities of the factors have been distribute by experts to construct the Bayesiannetwork model, which can simulate the project schedule risk. Finally, this study takesan actual case to prove the effectiveness of the proposed method and model.