Multi-objective optimization of the textile manufacturing process using deep-Q-network based multi-agent reinforcement learning