A Multi-agent Reinforcement Learning Model for Service Composition | IEEE Conference Publication | IEEE Xplore