complexity-econ/paper-05-endogenous: Endogenous Technology and Network Dynamics in AI Adoption

Paper-05: A Factorial Agent-Based Study This release contains the complete replication package for Paper-05 in the complexity-econ series. What's inside 10,080 Monte Carlo simulations across three campaigns: C1 Factorial: 2×2 design (fixed/endo σ × static/dynamic network) — 2,520 runs C2 Lambda sensitivity: 6 learning rates × 21 BDP levels — 3,780 runs C3 Rho sensitivity: 6 rewiring rates × 21 BDP levels — 3,780 runs 12 publication-ready figures (200 DPI) 7 Python analysis scripts generating all figures from raw CSV data LaTeX paper (17 pages, XeLaTeX + biblatex) with compiled PDF Key findings The reentrant (inverted-U) adoption shape survives endogenization in all four factorial cells Critical point BDP_c shifts by at most 250 PLN (from 500 to 750) under dynamic rewiring Combined effect is superadditive: +6.5 pp peak adoption vs static baseline Endogenous σ does not produce self-organized criticality (SOC) Non-monotonic ρ effect: moderate rewiring shifts BDP_c, high rewiring restores it Engine Requires complexity-econ/core v1.5+ with SIGMA_LAMBDA and REWIRE_RHO environment variable support.