A quasi-oppositional learning of updating quantum state and Q-learning based on the dung beetle algorithm for global optimization