Landscape Complexity Shapes the Role of Network Density and Diversity in Collective Adaptation Under Disruption

The polycrisis presents a future characterized by multiple co-occurring crises, demanding human collectives capable of navigating complex challenges simultaneously. This requires understanding how human collectives should be structured to improve their capacity to find and maintain effective solutions despite ongoing and frequent disruptions. We model collective problem-solving using an NK landscape impacted by intermittent disruptions to examine how network connectivity and agent diversity influence performance under ongoing disruptions. We measure the capacity for different collective structures to find and maintain high-performing solutions through various disruption regimes (frequency of disruption, impact size, impact distribution). Results reveal that disruptions degrade collective cumulative performance in simple landscapes (K = 0) but enhance it in complex ones (K = 7). In complex landscapes, moderate connectivity (d ≈ 0.2–0.32) and high diversity maximize cumulative performance, with disruptions amplifying these benefits. Additionally, we observe that impact distribution across agent groups affects cumulative performance, with skewed distributions. Our analysis demonstrates that in complex landscapes, disruptions can actually improve collective problem-solving capacity, but only with appropriate collective structures. This suggests that effective collective design requires matching structure to both problem complexity and disruption environment rather than applying universal principles.