A Dynamic QoS-Aware Logistics Service Composition Algorithm Based on Social Network

Abstract

The public logistics platform aims to provide customers with end-to-end logistics services by finding and composing a huge quantity of web services from logistics service providers. But, traditional service composition required predefined business process so that its flexibility is far from satisfactory in the problem. Path planning can be a solution of finding a suitable business path during service composition, but the search space will increase dramatically with the growth of service quantity and is hard to get a result within a tolerable interaction time. In the context of big data, to quickly build a service path with the optimal global QoS has become a problem demanding prompt solution. Sociologists point out that companies prefer familiar partners in the commercial environment. Using this principle, a concept of partner circle is defined, which can significantly reduce the search space in path planning. Combining path planning with service composition, a PartnerFirst algorithm is presented based on the social network, which is the cooperation network of service providers here. Simulation experiment shows that the PartnerFirst algorithm outperforms current approaches over 10 times in efficiency, with just about 10% loss in QoS. The relationship between efficiency and service quantity of the PartnerFirst algorithm is nearly linear. It proves that using social network in dynamic service composition is efficient and effective.

A Top-K QoS-Optimal Service Composition Approach Based on Service Dependency Graph

“…At present, many information technologies, such as cloud computing, big data, Internet-of-things, mobile computing and artificial intelligence, are being widely applied (Hu, Li, Cheng, Yu, Wang, & Bie, 2016a;Hu et al, 2016b;Pan, Lei, & Zhang, 2018;Peng, Leung, & Zheng, 2018;Xu et al, 2018;Xu, Dou, Zhang, & Chen, 2016;Xing, Hu, Yu, & Jiang, 2017). At the same time, as an emerging technique, Web service composition is developing fast and has received significant interest in the past years (Alrifai, Risse, & Nejdl, 2012;Almulla, Almatori, & Yahyaoui, 2011;Atzori, Lera, & Morabito, 2010;Yu, Chen, Lin, & Wang, 2014;Benouaret, Benslimane, Hadjali, & Barhamgi, 2011;Bao, Zhao, Shen, & Chen, 2016;Deng, Huang, & Tan, 2014;Gong, Qi, & Xu, 2018;Huang, Jiang, & Hu, 2009;Huo, & Wang, 2016;Jiang, Hu, & Liu, 2014;Jiang, Zhang, & Huang, 2010;Jaeger, Rojec-Goldmann, & Muhl, 2004;Menasce, 2004;Oh, Lee, & Kumara, 2007;Qi, Zhang, Dou, & Ni, 2017;Tao, Zhao, Hu, & Zhou, 2008;Wagner, Klein, & Klopper, 2012;Wagner, Ishikawa, & Honiden, 2011;Yan, Cui, Qi, Xu, & Zhang, 2018;Yan, Xu, & Gu, 2009;Zheng, Ma, Lyu, & King, 2011). Several efforts have led to the development of platforms and languages to support composition and deployment of services (Sheng et al, 2014;Chafle et al, 2007;Xia et al,2015).…”

Experimental Analysis for Semantic based Large Scale Service Composition using Deep Learning

“…Yang Yu et al [10] proposed a service composition based on Partner First algorithm. This algorithm combines path planning with service composition, presented based on the social network.…”

A time-aware dynamic service quality prediction approach for services

“…How to efficiently select and recommend services that meet users needs amongst a large-scale services market has therefore become a tremendous challenge [1,2] . Nowadays, QoS is being widely used in service composition [3][4][5][6] , service selection [7][8][9] , and service recommendation systems [10][11][12][13] . However, it is not possible for a user to use all services to get their QoS values, so it is often necessary to select or recommend services in the absence of QoS information.…”