Historical Saturation and Over-Consolidation in a Discrete Geometric–Densitary Architecture Model: Dynamic Reconstruction, Multi-Model Selection, and Asymptotic Regime Transition

This technical note presents an exploratory numerical study of a discrete Geometric–Densitary Architecture (AGD) model with memory, described through a consolidation variable and a historical-state variable. A homogeneous campaign of eleven reconstructed states was analyzed. The geometric baseline was dynamically estimated as Cgeom≈17.8929C_{\mathrm{geom}} \approx 17.8929Cgeom≈17.8929. Ten states satisfying the operational identifiability criterion Chist>0.5C_{\mathrm{hist}} > 0.5Chist>0.5 were used to reconstruct and fit the effective historical resistance. Six phenomenological models were compared. A power-law form, Reff(H)=R0[1+(HHc)p],R_{\mathrm{eff}}(H) = R_0\left[1+\left(\frac{H}{H_c}\right)^p\right],Reff(H)=R0[1+(HcH)p], provided the best fit, with R0≈11.8439R_0 \approx 11.8439R0≈11.8439, Hc≈526.994H_c \approx 526.994Hc≈526.994, p≈1.60239p \approx 1.60239p≈1.60239, and RMSE ≈1.3204\approx 1.3204≈1.3204. The fitted relation predicts a maximum consolidation near H≈723.07H \approx 723.07H≈723.07, followed by a decreasing historical contribution interpreted operationally as an over-consolidation regime. The numerical campaign also shows a transition from runs classified as fixed-point convergence to runs requiring temporal averaging. This behavior is compatible with a change in asymptotic dynamical regime, but does not yet establish a bifurcation, limit cycle, phase transition, or dynamical hysteresis. The work is intentionally limited to a scalar exploratory model. It does not claim a physical property of the vacuum, a cosmological theory, or an alternative to dark matter. Spatial extensions, stability analysis, bidirectional parameter sweeps, and additional model-selection tests are identified as future work.