Research On A Simulation-based Dynamic Interaction Network For Project Risk Response Decisions

Due to the growing project complexity of technique and organization,the risk interaction arising from how a given risk may trigger numerous risks or resulting from others is gaining research attention and warrant further study.In order to address numerous risk interactions,a project risk management process considering the risk interaction is thus developed,comprising risk interaction network(RIN)modeling,RIN analysis,and RIN response.However,most of the research studies focus on the methods of RIN analysis and RIN response based on a static RIN,and very few studies consider the dynamic trait of risks and construct the dynamic RIN consisting of several static RINs to analyze and respond.While none of them characterize the delay-oriented risk propagation in the whole project life cycle,leading to inaccurate risk criticality assessment and hence invalid actions taken in the risk response phase.Thus,this paper considers the dynamic trait of project risks and the time delay in risk interaction and proposes methods to solve the following problems,i.e.,how to construct the dynamic risk interaction network(DRIN)and how to make risk response decisions based on such DRIN.First,this paper integrates the occurrence time of risks and the time delay in risk interactions as the attributes of network elements and proposes effective methods to identify and assess DRIN,which is composed of nodes(dynamic risks)and edges(risk interactions).Specifically,to identify DRIN nodes,work breakdown structure and risk breakdown structure are introduced to generate a coupling matrix.To assess DRIN node attributes,the Delphi-based method and set-valued statistical method are used to quantify the spontaneous probability and loss due to a given risk respectively,and a method regarding risk occurrence time assessment is proposed.Experts are invited to identify DRIN edges(risk interactions)using risk structure matrix and analytic hierarchy process is adopted to quantify the likelihood of triggering one risk by another,i.e.,the transition probability,and a method to quantify the time delay is proposed.Subsequently,two types of risk response decisions plan,i.e.,node-oriented plans and edgeoriented plans,are designed based on the constructed DRIN.A project is prone to generate economic loss if the risks are not well taken care of.Reducing total risk loss is thus served as the target of the risk response decision.The calculation of risk criticality and total risk loss relies on probability of risk occurrence,while traditional methods are insufficient to compute the accurate occurrence likelihood of a given risk due to risk interactions.For this,a simulationbased network algorithm is used to compute the frequency of every risk’s occurrence in the project lifecycle,to proxy the probability of risk occurrence,and then calculate total risk loss.After ascertaining the target of risk response decisions,two types of risk response decisions plans are designed for two network elements respectively,i.e.,node-oriented plans and edgeoriented plans.From the perspective of eliminating key nodes(risks),the former type selects the network elements which can occur the key risks through paths in DRIN to obtain the subnet of critical risk,devises a network index to evaluate the contribution of each element to the critical risk occurrence through subnet structure analysis,and then proposes the basic risk response strategy.As risk response decision is a set of risk response strategies under limited budget constraints,two specific node-oriented plans are formed according to the basic strategy,i.e.,direct response plan and iterative response plan.Similarly,another novel network index is designed for edge-oriented plans to evaluate the significance of each edge through network structure analysis proposes another basic strategy,and form a direct response plan and an iterative response plan.Finally,based on the constructed DRIN and the proposed risk response decision plans,we design the algorithm and adopt the simulation method to verify the necessity of applying DRIN to manage project risks and ascertain applicable scenarios of each risk response decision plan,supporting project managers to make risk response decisions.By introducing the methods of dynamic risks identification and assessment as well as the time delay assessment,we expand the existing research from static risk interaction network to dynamic risk interaction network,which solves the problem that the existing research are incapable to describe the characteristics of project risks and risks interactions.Distinguished from the existing static network index,novel network indices suitable for the dynamic network are designed by adding a time parameter,which can provide ideas for dynamic network analysis in other fields.Clarifying the application scenarios of each risk response decision plans with the help of simulation,provides systematic risk response decision methods for project managers.