Uncovering configurational pathways of freeway diverge congestion via integrated fsQCA and PLS-SEM

Suburban freeway diverge areas are recurrent bottlenecks where congestion results from non-linear interactions. Adopting a configurational perspective, this study integrates fsQCA and PLS-SEM to identify and quantify configurational pathways associated with congestion. Empirical data from six interchanges indicate that frequent mandatory lane-changing is a necessary condition, while no single factor is sufficient. Four sufficient configurations are identified: (1) high heavy-vehicle proportion with frequent mandatory lane-changing; (2) high traffic volume with frequent mandatory lane-changing; (3) frequent mandatory lane-changing, single-lane ramp, and rainfall; and (4) high traffic volume, high diversion ratio, high heavy-vehicle proportion, single-lane ramp, and rainfall. PLS-SEM results indicate that configurations combining high traffic volume and frequent mandatory lane-changing show the strongest association. Normalised importance analysis ranks mandatory lane-changing (0.389) highest, followed by heavy-vehicle proportion (0.225), traffic volume (0.141), and ramp lane type (0.124). This framework provides a configuration-centred account of congestion formation and supports targeted mitigation at freeway diverges.