GenAI-enhanced risk response in project portfolios: a complex network evolutionary game approach
研究了生成式人工智能如何通过整合多模态项目数据、协调管理者理解来促进项目组合中的协同风险应对行为，并利用复杂网络演化博弈模型分析了其扩散机制。
Project portfolios increasingly face complex, interdependent risks that challenge operational resilience. Effective Project Portfolio Risk Response Decisions (PPRRDs) require Synergistic Response Behaviours (SRBs) among project managers, yet SRB adoption is often impeded by interpretational heterogeneity and incomplete risk information. Generative Artificial Intelligence (GenAI) offers new opportunities to improve coordination by integrating multimodal project data, harmonising managerial interpretations, and enhancing local risk response capacity. This study proposes a GenAI-enhanced PPRRD framework and develops an evolutionary game formulation to examine SRB dynamics under both GenAI-supported and conventional conditions. The two-project game model is extended to a complex network evolutionary game (CNEG) to analyse SRB diffusion across portfolio networks. A simulation based on a large-scale construction project portfolio illustrates how network topology, project heterogeneity, and GenAI-related parameters shape equilibrium outcomes. The findings show that (i) GenAI accelerates SRB diffusion and increases portfolio-level risk response payoffs; (ii) reward–penalty mechanisms substantially amplify GenAI’s effects on diffusion; and (iii) portfolios with inherently strong risk response capabilities exhibit slower SRB diffusion. The study offers a novel analytical foundation for integrating GenAI into project portfolio risk governance to enhance operational resilience.