FUGNN: Harmonizing Fairness and Utility in Graph Neural Networks