A profit optimization oriented service selection method for dynamic service composition

Service composition is an effective means of building value-added service in service-oriented computing environment. Current research focuses on the fulfillment of customer value, while neglects the value procurable by service broker, which is the compositor of individual services as well as the provider of composite services. On the one hand, over-optimized service quality will not bring additional profit to the service provider as well as no remarkable improvement to customer satisfaction, thus is unnecessary for the value of both sides of service participants in SLA environment; on the other hand, due to the uncertainty of both services and the environment for delivering services, real quality of service-oriented applications exhibits as uncertain, too. So real services may not meet the quality requirement of negotiated service level, or even fail. Profit and service strategies are studied for SLA, and a novel service selection model is proposed for profit optimization. Based on periodical estimation of service cost and instant feedbacks, service requirements are greedily scheduled and optimized service selection is realized for dynamic service composition based on simulated annealing algorithm. Experimental results show that this approach does not only promote the profit of composite services, but also have superior efficiency in procuring optimized results under different circumstances of requirements distribution, compared with traditional approaches.