A probabilistic approach to assessing project complexity dynamics under uncertainty

Intractable complexity may be encountered as the construction project advances. Existing research rarely investigates the time-updated dynamic in project complexity as the project process progresses. This study develops a novel systematic soft computing approach based on Bayesian inference to explore the evolutionary dynamics in project complexity under uncertainty. By learning the network structure and parameters from given data, a dynamic Bayesian network model is established to simulate the complex interrelations among 7 complexity-related variables. The developed approach is capable of performing predictive, sensitivity, and diagnostic analysis on a quantitative basis. The construction project of EXPO 2010 is used to testify the effectiveness and applicability of the developed approach. Results indicate that (1) more attention should be paid to technological complexity and task complexity in the process of complexity management; (2) the developed dynamic Bayesian network approach can model the evolutionary dynamics of project complexity at different scenarios; and (3) the complexity level of a specific construction project over time can be predicted in a dynamic manner. This research contributes to (a) the state of the knowledge by proposing a systematic soft computing methodology that can model and identify the dynamic interactions of project complexity factors over time, and (b) the state of the practice by gaining a better understanding of the most sensitive factors for managing complexity in a changing project environment.