Optimal reinsurance with model uncertainty and Stackelberg game