Track: Decision Sciences

Abstract

The growing complexity of contemporary projects increases the chance that project risks with different natures are interrelated through complex cause-effect relationships. However, the effects of risk interdependencies and dynamic risk propagations are often ignored in previous project risk management studies, leading to inconsistent risk assessment results with engineering practice and reduced efficacy of risk treatment actions. This research proposes an intelligent decision support system (DSS) to determine an effective portfolio of risk treatment actions by considering complex risk interdependencies and project resource constraints. First, a hierarchical project risk interdependency network (RIN) is developed systematically based on the interpretive structural modelling process, where all possible cause-effect interdependencies among project risks are presented. Then, a Monte Carlo simulation (MCS)-based RIN model is devised to prioritise project risks and evaluate the overall project risk level, where the stochastic behaviour of risk occurrence and risk propagation effects within a project RIN are captured. Subsequently, a bi-objective risk treatment optimisation model is proposed considering limited project resources, aiming to minimise not only the project loss due to risk but also the risk treatment cost. To obtain a set of Pareto-optimal risk treatment solutions, the non-dominated sorting genetic algorithm II is tailored by integrating with the MCS-based RIN model. A case study is provided to demonstrate the feasibility of the proposed DSS for project risk treatment decision-making. The findings from this work help project decision-makers allocate risk treatment budgets more appropriately and determine the most effective risk treatment solution corresponding to their risk attitudes.