Integrating reinforcement learning and skyline computing for adaptive service composition

Wang, Hongbing; Hu, Xingguo; Yu, Qi; Gu, Mingzhu; Zhao, Wei; Jia, Yan; Hong, Tianjing

doi:10.1016/j.ins.2020.01.039

Cited by 31 publications

(23 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most popular are genetic algorithms that are used to optimize business processes during runtime. Further advances in adaptive service composition that fulfill user requirements were shown by Wang et al [22]. Another important aspect, the automated discovery of crossorganizational collaborative business processes was shown by Montarnal et al [23].…”

Section: Related Workmentioning

confidence: 99%

Integration Flows Modeling in the Context of Architectural Views

Górski

2023

IEEE Access

View full text Add to dashboard Cite

In an increasing number of software applications, the execution of their functions depends on communication with external systems. Messaging enables the integration of information technology (IT) systems in a loosely-coupled manner. The paper aims to show the Integrated services architectural view and two methods of modeling messaging flows at the service and business levels. Service flows allow for modeling the communication between systems/services at the level of sending messages. In contrast, business flows model the entire course of interaction among cooperating applications. The methods employ an Integration flows diagram, which is a specialized version of the Unified Modeling Language (UML) Activity diagram. In addition, the business process task from the Integrated Processes view defines the context for the business flow. Moreover, the Use Cases view identifies the integration scope within requirements. All those views belong to the 1+5 architectural views model. The paper exerts extensibility mechanisms declared in the UML Profile for Messaging Patterns. Besides, the profile has been augmented with stereotypes for messaging patterns for the Apache Camel framework, stakeholders involved in communication, and up-to-date stereotypes for structural components. The methods were used in the integration design of sending orders and confirmations between the business applications of the brokerage house and the stock exchange.

show abstract

Section: Related Workmentioning

confidence: 99%

Integration Flows Modeling in the Context of Architectural Views

Górski

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…There have been some attempts to apply reinforcement learning to the composition of web services. A series of works utilized multiple reinforcement learning agents to search for optimal services in terms of QoS for a given web service composition problem [25], [26]. At each state, reinforcement learning agents select one of the candidate services for the corresponding function and transit to one of the next states until it reaches a terminal state.…”

Section: Related Workmentioning

confidence: 99%

“…Searching for the best combination of the services is a challenging problem because of the large set of candidate services, correlated QoS of candidate services [13], and the uncertainties of reality [27]. Therefore, traditional approaches to service composition mainly focus on reducing the search space to improve the efficiency of algorithms [26], [29]. By contrast, the main focus of our approach is on the generalization to learn an optimal policy for selecting services that can be applied to various situations generally and can adapt to real-time variations in the environments.…”

Section: Related Workmentioning

confidence: 99%

Dynamic and Effect-Driven Output Service Selection for IoT Environments Using Deep Reinforcement Learning

Baek

2023

IEEE Internet Things J.

View full text Add to dashboard Cite

In the context of the recent emergence of the internet of things (IoT), human users and IoT-based services are interacting via physical effects such as light and sound. Therefore, it is necessary to consider the quality of the delivery of physical effects to users by IoT devices for selecting services in IoT environments. However, traditional service-selection algorithms focus primarily on the network-level quality of service (QoS), such as latency and throughput. In this study, we improve on the visual-service effectiveness metric developed in our previous work to measure the effectiveness of the personalized delivery of physical effects of visual services to users by considering user-and application-specific factors. We evaluate the metric by conducting a user study, and the results show that the metric reflects users' perceived effectiveness with high accuracy. We also investigate the use of virtual reality (VR) to imitate physical environments for efficient evaluation of the metric. Based on this metric, we develop a dynamic effect-driven output-service selection agent (DEOSA) that selects output services dynamically by considering the effectiveness of service-effect delivery. By adopting a stateof-the-art reinforcement-learning algorithm, DEOSA can learn the optimal policy for selecting output services that can be generalized to various environments. We evaluate DEOSA in simulated IoT environments and show that it can learn the optimal policy successfully; it generally performs better than traditional greedy algorithms in terms of the visual service effectiveness metric and the replacement overhead in randomlygenerated test environments.

show abstract

“…As the IaaS composition is dynamic in nature, a model-free learning platform might be more suitable than a model-based learning platform [15]. Adaptive service composition utilizes reinforcement learning based methods to compose services from a consumer's perspective [62]. Markov Decision Process or MDP is often used to model service composition problem [1].…”

Section: Related Workmentioning

confidence: 99%

Sequential Learning-based IaaS Composition

2021

View full text Add to dashboard Cite

We propose a novel Infrastructure-as-a-Service composition framework that selects an optimal set of consumer requests according to the provider’s qualitative preferences on long-term service provisions. Decision variables are included in the temporal conditional preference networks to represent qualitative preferences for both short-term and long-term consumers. The global preference ranking of a set of requests is computed using a k -d tree indexing-based temporal similarity measure approach. We propose an extended three-dimensional Q-learning approach to maximize the global preference ranking. We design the on-policy-based sequential selection learning approach that applies the length of request to accept or reject requests in a composition. The proposed on-policy-based learning method reuses historical experiences or policies of sequential optimization using an agglomerative clustering approach. Experimental results prove the feasibility of the proposed framework.

show abstract

Integrating reinforcement learning and skyline computing for adaptive service composition

Cited by 31 publications

References 14 publications

Integration Flows Modeling in the Context of Architectural Views

Integration Flows Modeling in the Context of Architectural Views

Dynamic and Effect-Driven Output Service Selection for IoT Environments Using Deep Reinforcement Learning

Sequential Learning-based IaaS Composition

Contact Info

Product

Resources

About