Continuous Growth Mechanism Of Service Network

With an increase in the number of the Internet-based services, more and more people choose to meet their life needs via the Internet. Service network is an effective way to solve the challenge of mass customization of services. Compared with traditional approaches, service network it reduces the cost of searching in massive candidate services for customers and improves the cost effectiveness of service offerings for service providers. However, customer requirements in terms of the functionality and Qo S tend to being more and more diversifed and individualized, and the common characteristic of massive customer requirements evolve over time. In such circumstance, an existing service network may not have the capacity of fully meeting the subsequent arriving requirements, and consequently, the customer satisfaction degree would be reduced; or, some services in the service network are not or less involved in meeting requirements, which results in resource waste of service providers. Therefore, a service network should not keep static but need to dynamically and continuously grow in order to adapt to the requirement changes. This is what is called the problem of service network dynamic continuous growth.In traditional service computing research, the study includes service replacement, service recombination, service system evolution, etc which is aimed at adapting to changing in requirement. However, these methods focus on aspects of repairment of the software system and interface protocols, and often only for a single service composition of a customer requirement. Therefore, dealing with the larger group customer requirements evolution problems has a high price with these methods.To solve above problems, the paper proposes a method of service network dynamic continuous growth that gives a service network growth plan and optimized solutions of customization of requirements meanwhile. The input of the problem of service network dynamic continuous growth is customer requirements and a service network. This paper grows the service network specific to the inadequate capacity founded in the process of service network customization, and assesses the cost effectiveness. The final output includes growth plan of the service network and customization plan of requirements. The main research contents include:(1) The inadequate capacity in the process of service network customization are classified, several kinds of typical mechanism of the service network dynamic continuous growth are explored and an indicator of assessing service network dynamic gains and losses is given which will be the optimization goal of the service network dynamic continuous growth optimization algorithm.(2) Propose service network growth optimization algorithm specific to multi requirements arrived at the same time. It refers that the service network doesn’t grow for individual requirement, but needs to wait for the follow-up requirements and make decision for service network growth at the same time for n new requirements, considering that the n requirements are concurrent. First of all, establish a mathematical model of the algorithm. The algorithm aims to give the optimal service network growth plan with the optimal gains and losses assessment standards and the optimal customized plans which meet the requirements in terms of the Qo S and functionality with the input of n requirements and a service network. At last, through the experiment, verify the validity of the algorithm in terms of cost and time by being compared with the service network construction algorithm for multi requirements, and verify the influence factors of algorithm.(3) Characteristics analysis was carried out on the accumulated historical requirements data to get the relationship between requirements characteristic and customization plan of requirements. Based on this, the service network customization and growth algorithm based on the requirements characteristics is proposed. The algorithm predicts the customization plan of requirements and growth result of service network with the input of n concurrent requirements that arrived at the same time at the current moment and a service network by analyzing the historical results of customization and growth plans of service network, under this premise of not carry on the concrete custom. Present mathematical model and solution strategy of the algorithm and through the experiment, contrast with demand customization method based on ABC and service network growth optimization algorithm specific to the multi requirements arrived at the same time to validate the superiority of customization plan of the requirements and the time cost effectiveness of the growth of the service network respectively, and verify the influence factors of algorithm.(4) The environment supporting the deployment of phased growth of the service network is developed, which sustains the deployment of multi-version of the continuous growing service network and the visualization of all functions in the system above.