Optimizing selection of competing services with probabilistic hierarchical refinement

Tan, Tian; Chen, Manman; Sun, Jun; Liu, Yang; André, Étienne; Xue, Yinxing; Dong, Jin Song

doi:10.1145/2884781.2884861

Cited by 21 publications

(11 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It can be applied to resource-critical systems and domains that seek to make the best use of available resources and can tolerate some degradation or deviation of the requirements. A travel agency example used to illustrate composition [40] as a holiday package consists of a hotel, a flight, and a car. When the requirements cannot be satisfied, several possible adaptations can be made to assist the user in making their decision, rather than letting the users test and see.…”

Section: Discussionmentioning

confidence: 99%

“…For systems with more than a few components the search space is unmanageable, large and complex. In [40] the authors propose to select a concrete service for each abstract service defined within a service composition, in order to optimise overall quality of service. However, this solution does not enable reasoning about requirements at runtime.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Won't Take No for an Answer: Resource-Driven Requirements Adaptation

Bennaceur

Zisman

McCormick

et al. 2019

2019 IEEE/ACM 14th International Symposium on Software Engineering for Adaptive and Self-Managing Systems (SEAMS)

View full text Add to dashboard Cite

Adaptive composition dynamically and opportunistically uses and combines resources to best satisfy user requirements. However, when available resources cannot satisfy those requirements, no guidance or alternative options are offered by existing composition solutions. In this paper we address this issue by presenting an approach that tries to find substitutions for unavailable resources while satisfying the initial requirements. If no satisfactory substitutions are found, the requirements are adapted based on the resources available. Given that such requirements adaptation might be unbounded, we limit the search space guided by the available resources. Our approach ensures that alternative compositions given to users are achievable using available resources. We demonstrate the validity of our approach by implementing a prototype tool and applying it to support individuals in meal planning to reduce food waste.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Won't Take No for an Answer: Resource-Driven Requirements Adaptation

Bennaceur

Zisman

McCormick

et al. 2019

2019 IEEE/ACM 14th International Symposium on Software Engineering for Adaptive and Self-Managing Systems (SEAMS)

View full text Add to dashboard Cite

show abstract

“…Then each execution path is extracted and the optimal global solution is obtained by solving the mixed integer linear programming problem. Tan et al [5] reduce the search space based on local optimality of the services and global constraints. However, these approaches would cost too long when the candidate services set becomes larger.…”

Section: B Qos-aware Service Compositionmentioning

confidence: 99%

“…This way of finding the best combination would take a lot of time and resources. As web service composition needs to be handled at runtime, users are more willing to accept a near-optimal combination with lower cost, rather than an exact one with higher cost [5].…”

Section: Problem Descriptionmentioning

confidence: 99%

“…As illustrated in Fig.1, there exists a business process with 5 tasks, and for each task there are 4 alternative web services with similar functionality but different QoS values. Then the number of alternative composite services will be 4 5 . If the number of tasks or the number of alternative web services for each task increases, the number of alternative composite services grows exponentially.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Clusters Adaptive Brain Storm Optimization Algorithm for QoS-Aware Service Composition

Peng

Wang

2020

IEEE Access

View full text Add to dashboard Cite

Service composition provides an effective means to fulfill users' personalized requirements and complex business applications. With the number of web services rapidly growing, finding the best combination among services based on quality of service (QoS) poses a critical computational challenge due to the exponential growth of alternative composite services. Some efforts are developed to find the near-optimal service combination within an acceptable time range. They fall into two categories of solutions: exploring partial combinations in the service space and downsizing the optimization problem in scale. Although they solve the scalability problem to some extent, the required computational time is usually high. A promising direction is to integrate these two categories of solutions. However, its practical application suffers from three challenges: no good search scheme, no consideration of the natural organization of sub-problems, and the lack of diverse combinations. In this work, we propose a novel approach, called multi-clusters adaptive brain storm optimization (MCaBSO) algorithm. The proposed method uses brain storm optimization (BSO) as the search scheme to combine the division of search space with the exploration of the reduced search space. MCaBSO uses the twin support vector machine (TWSVM) to effectively divide the search space according to the natural organization of sub-problems. MCaBSO provides an adaptive dual strategy that gives guidance for the generation of diverse combinations. MCaBSO enables the agile exploration of the reduced search space and generates more high-quality combinations. MCaBSO is evaluated on two datasets to show effectiveness and efficiency.

show abstract