Multi-agent Web Service Composition using Partially Observable Markov Decision Process

Abstract

Web Service Composition using Markov Decision Process provides a good result using Reinforcement Learning of Artificial Intelligence. But the problem when dealing with MDP is that is assumes that all the information related to the variable is provided accurately. In case of QoS aware Web Service Composition it is not possible that all the QoS parameters are accurately provided. So to handle legacy of QoS parameter and still doing Web Service Composition is an issue. This paper provides a technique to deal with this problem.