Paper Info
Title
Probability Matrix of Request-Solution Mapping for Efficient Service Selection
Abstract
With more and more Web services flooded on the Internet, the scale of Web services and complexity of connections among them are growing rapidly. This phenomenon has brought great challenges to service selection. Due to the huge search space, existing research approaches are hardly feasible in dynamic real-time scenarios under a stringent time limit with a large number of potential Web services involved. In order to deal with this problem, the focus of this paper is to improve the efficiency of QoS-aware web service selection in real-time considering a priori knowledge from historical log, which can reduce the search space effectively. We first analyse and discover the distribution of customer requests to identify request clusters, and we mine valuable fragments or service patterns from historical service solutions. Then, we build a probability matrix to improve the efficiency of service selection algorithm, which contains the request-solution mapping relationships between request clusters and service patterns based on statistical method. A series of experiments using both real and synthetic data demonstrate that our approach improves Global Planning optimisation algorithm (GP) and Artificial Bee Colony algorithm (ABC) by 36.20% and 41.98% respectively.
Year
DOI
Venue
2017
10.1109/ICWS.2017.51
2017 IEEE International Conference on Web Services (ICWS)
Keywords
Field
DocType
service pattern,request clustering,probability matrix,request-solution mapping
Data mining,Artificial bee colony algorithm,Computer science,A priori and a posteriori,Quality of service,Synthetic data,Knowledge engineering,Cluster analysis,Web service,The Internet
Conference
ISBN
Citations 
PageRank 
978-1-5386-0753-4
1
0.35
References 
Authors
20
5
Name
Order
Citations
PageRank
Ruilin Liu115917.17
Xiaofei Xu240870.26
Zhong-Jie Wang335664.60
Quan Z. Sheng43520301.77
Hanchuan Xu5147.04