Dynamic Network Analysis of the Risks of Mega Infrastructure Projects from a Sustainable Development Perspective

Abstract

Mega infrastructure projects (MIPs) are exposed to numerous interdependent risks of various natures which pose difficulties in risk management. Thus far, the research on the risk interactions of MIPs has been focused on developing static risk networks within a single category of risks, at certain stages of the project. It is essential to understand the risk interactions at various stages of MIPs to identify the key risks and key risk relationships that jeopardise their success. This is especially relevant nowadays, as MIPs are expected to be delivered sustainably. Therefore, to analyse the dynamic risk interaction of MIPs, initially, through literature analysis and expert interviews, combined with the four dimensions of sustainable development and the four stages of MIPs, 98 risk factors of MIPs were identified. Subsequently, semi-structured interviews were conducted to determine risk relationships and weights. Risk networks were developed for each stage of MIPs, and improved social network analysis was applied to these risk networks. Finally, the key risks and key risk relationships in each stage of MIPs were identified by analysing the changes of multi-level network indicators. This aided in determining risk control strategies. The results demonstrate that the key risks and key risk relationships are different for each stage of MIPs. Furthermore, the risks of different dimensions of sustainable development have different relationships at different stages. This research is the first to identify the risk relationships involved in MIPs by taking into consideration the whole project life cycle and its sustainable development. This research provides theoretical support for the risk management of MIPs, and strategic suggestions for controlling the risks at each stage of the project.

Interaction mechanisms of interface management risks in complex systems of high-speed rail construction projects: an association rule mining-based modeling framework

“…On the other hand, scientifically extracting the correlations between factors and their strengths and constructing the interface management risk interaction network is the basis for revealing the interface management risk interaction mechanism. However, existing risk interaction network modeling methods rely heavily on the experience of the modeler or domain experts, which is accompanied by lower modeling efficiency and greater subjectivity and may lead to lower reliability of the analysis results (Li et al ., 2022). For example, in this paper, 91 IMRs exist in a high-speed rail construction project.…”

“…Complex networks use nodes to represent elements and their characteristics in a complex system and edges to describe their relationships. In weighted networks, the weights of edges are related to their strengths (Li et al ., 2022; Shi et al ., 2022). This study uses ARM to improve the modeling method of weighted complex networks.…”

“…GIGI represents the global influence of a node group on another node group based on the sum of links originating from the reciprocal of reachability and geo-distance from group n to group m . Detailed information and formulas for these indicators can be found in Li et al ., (2022) and Wang et al ., (2017).…”

“…They are modified from the corresponding indicators of an unweighted network by introducing a weighting parameter α to consider the weight of relationships (Table 2) (Opsahl et al, 2010). Li et al suggested setting α to 0.5 (Li et al, 2022).…”

“…Similarly, in-RRGD indicates the global impact of other nodes on a particular node in the network. Additionally, DIGI represents the direct influence of a node group on another node group based on the number of connections between group n and group m. GIGI represents the global influence of a node group on another node group based on the sum of links originating from the reciprocal of reachability and geo-distance from group n to group m. Detailed information and formulas for these indicators can be found in Li et al, (2022) and Wang et al, (2017).…”