Sequential Learning-based IaaS Composition

We propose a novel Infrastructure-as-a-Service composition framework that selects an optimal set of consumer requests according to the provider’s qualitative preferences on long-term service provisions. Decision variables are included in temporal conditional preference networks represent for both short-term and consumers. The global ranking is computed using k -d tree indexing-based similarity measure approach. extended three-dimensional Q-learning approach maximize ranking. design on-policy-based sequential selection learning applies length request accept or reject composition. proposed method reuses historical experiences policies optimization agglomerative clustering Experimental results prove feasibility framework.