A Cost-Effective Service Composition Method for Mass Customized QoS Requirements

Abstract

We present a cost-effective service composition algorithm MC4MR aiming at mass customized QoS requirements. If a group of customers raise personalized QoS constraints on a given service, the algorithm looks for a finite number of composition solutions to meet these requirements and realize maximum benefit. To pursue higher cost-effectiveness, we combine multiple individualized requirements together and find a limited number of composition solutions that jointly satisfy them. The algorithm adopts the recursive and greedy strategies to reduce the computation complexity. Mass requirements are ordered according to the "potential benefit (PB)", requirements with higher PB are handled earlier, and subsequent ones are handled based on previously obtained solutions using three heuristic policies. The MC4MR algorithm is further simplified as an OC4MR algorithm (to look for one solution for multiple customized requirements) and an OC4OR algorithm (to look for one solution for one requirement). Experiments are conducted to compare performance and cost-effectiveness between MC4MR and traditional approaches, and some factors that impact the quality of solutions are explored.