Category: Network Sci.

Summary: Testing when adaptive sanctions help sustain cooperation in a spatial public-goods game and when long delays, noisy sensing, and enforcement fatigue make them backfire.

Adaptive punishment can in principle stabilize cooperation by responding to local defection, but real intervention systems are delayed, noisy, and costly to sustain. This experiment asks when those frictions become strong enough that adaptive sanctions perform worse than a simple fixed policy in a spatial public-goods game.

The model simulates cooperation on a lattice under several observation delays, noise levels, fatigue rates, and synergy strengths, then compares adaptive and fixed sanction strategies. It tracks cooperation levels, collapse fractions, and the extent to which growing enforcement fatigue erodes the benefit of adaptive control.

That turns a standard evolutionary-game setup into a delayed-control problem. The main scientific question is whether there is a regime inversion in which feedback that helps under fast accurate observation becomes harmful once the same system responds too slowly and tires out.

Method: Lattice public-goods simulations comparing adaptive and fixed sanctions across observation delay, noise, fatigue, and synergy conditions.

What is measured: Final cooperation level, collapse fraction, fatigue level, adaptive-minus-fixed cooperation difference, delay-driven regime inversion, and condition-by-condition policy contrasts.