Megaprojects with increasing scale and complexity necessitate collaborative innovation, which is marked by dynamic evolution. Networks are an effective technique for analyzing interorganizational interactions, yet few studies have explored the complex mechanisms driving the evolution of collaborative innovation networks (CINs) in megaprojects. This study constructed a CIN by longitudinally tracking patents from the Derwent Innovation Index database. Social network analysis and network motif analysis were employed to capture the evolutionary characteristics, and the stochastic actor–oriented model was introduced to identify the underlying driving forces behind the evolution. Our results corroborated that the CIN exhibited scale-free characteristics, evolving from star-centralized to connected-distributed structures. Triadic closure facilitated the evolution of collaborative innovation. Similar innovation abilities and experiences fostered collaboration, whereas similar innovation influences exerted a negative impact. Geographical, cognitive, and organizational proximity all acted as drivers; however, institutional proximity was not significant. These findings further deepen our understanding of the dynamics of collaboration innovation in megaprojects and offer valuable insights for formulating strategies to enhance interorganizational collaboration.
Ning et al. (Sun,) studied this question.