Authors: Wanchun Yang; Chenxi Zhang; Bin Mu
Addresses: School of Electronics and Information Engineering, Tongji University, Shanghai 201804, China ' School of Software Engineering, Tongji University, Shanghai 201804, China ' School of Software Engineering, Tongji University, Shanghai 201804, China
Abstract: With the growing number of services being deployed in cloud environment, it becomes difficult to facilitate mashup quickly. In this paper, we introduce a comprehensive evaluation model of mashup which considers three aspects, including functional similarity, QoS characteristics and transactional properties, and use them together as optimisation criteria. To enlarge the scope of feasible solutions, we adopt Analytic Hierarchy Process (AHP) to analyse the global constraints and consider relevant services with different granularity. An efficient mashup optimisation algorithm based on skyline and hybrid particle swarm optimisation is proposed to seek a near-to-optimal solution. To overcome the premature convergence of traditional particle swarm optimisation, a series of effective strategies are presented, which include global mutation operation and adaptive inertia weight parameter. The experimental results show that our proposed approach is feasible and effective.
Keywords: mashup optimisation; cloud computing; global constraints; granularity; skyline; hybrid PSO; particle swarm optimisation; functional similarity; QoS characteristics; quality of service; transactional properties; analytical hierarchy process; AHP.
International Journal of Grid and Utility Computing, 2014 Vol.5 No.4, pp.227 - 235
Received: 02 Sep 2013
Accepted: 15 Dec 2013
Published online: 09 May 2015 *