Beyond Monte Carlo: leveraging temporal difference learning for superior performance in dynamic resource allocation Konferenz-Paper