QoS Aware Semantic Web Service Composition Approach Considering Pre/Postconditions

Abstract

Web service composition is a topic bringing several issues to be resolved. Our work deals with the effectiveness and scalability of service composition. During composition we consider QoS and pre-/post-conditions of single services to create a composite service satisfying the user needs the best. Regarding pre-/post-conditions we propose an approach to fast determination of which services produce results expected by the user, i.e. the post-condition of which services implicates the desired condition defined in the user goal. This paper presents also a renewed version of our approach to QoS aware service composition. We achieved a dramatic improvement in terms of composition time by performing a useful restriction on the service space.

Continuous Query for QoS-Aware Automatic Service Composition

“…Inspired by Dijkstra algorithm [19], we make use of a similar order to update the status of affected services. The distance changes 4 in Dijkstra algorithm are +∞ → X (0 < X < +∞). In our problem, for the first category of affected services whose allQoS got improved, we can handle it like Dijkstra algorithm; For the second category, we transform it into the first category by assigning the original allQoS of these services to +∞, while its new allQoS is not reassigned.…”

“…A final affected service means its allQoS is different from the original value after all affections are processed. 4 Let Vs be the start node, the distance of every node is the distance from Vs to it. Then, a forward search is conducted from Vs, all the reachable nodes are inserted into a priority queue with their distances.…”

“…With the consideration of non-functional preferences of users and mass services available nowadays, QoS-aware automatic service composition has attracted a lot of attraction [3], [4], [5], [6]. Web service research community has hosted Web Service Challenge [16] to solicit software for the problem.…”

“…No predefined process template is needed in this problem. Furthermore, to guarantee the overall Quality-of-Service (QoS) of composite services, different composition techniques have been studied in [3], [4], [5], [6] to generate composition service with favorite QoS as well as correct function. By and large, all of the aforementioned approaches share a common assumption such that characteristics and behaviors of existing Web services seldom change.…”

An approach towards adaptive service composition in markets of composed services

“…The majority of existing composition approaches can be considered as symbolic techniques that base on explicitly given information [5][6][7][8][9][10][11][12][13][14][15]. Alternatively, machine learning techniques are proposed to replace symbolic techniques (cf.…”