Using Ensemble and TOPSIS with AHP for Classification and Selection of Web Services

Due to the increasing number of Web Services with the same functionality, selecting a Web Service that best serves the needs of the Web Client has become a tremendously challenging task. Present approaches use non-functional parameters of the Web Services but they do not consider any preprocessing of the set of functionally Similar Web Services. The lack of preprocessing results in increased use of computational resources due to unnecessary processing of Web Services that have a very low to no chance of satisfying the consumer’s requirements. In this paper, we propose an Ensemble classification method for preprocessing and a Web Service Selection method based on the Quality of Service (QoS) parameters. Once the most eligible Web Services are enumerated through classification, they are ranked using the Technique of Order Preference by Similarity to Ideal Solution (TOPSIS) method with Analytic Hierarchy Process (AHP) used for weight calculation. A prototype of the method is developed, and experiments are conducted on a real-world Web Services dataset. Results demonstrate the feasibility of the proposed method.

QoS aware web service selection using orthogonal array learning on fruit fly optimization approach

Purpose This paper aims to solve the web service selection problem using an efficient meta-heuristic algorithm. The problem of selecting a set of web services from a large-scale service environment (web service repository) while maintaining Quality-of-Service (QoS), is referred to as web service selection (WSS). With the explosive growth of internet services, managing and selecting the proper services (or say web service) has become a pertinent research issue. Design/methodology/approach In this paper, to address WSS problem, the authors propose a new modified fruit fly optimization approach, called orthogonal array-based learning in fruit fly optimizer (OL-FOA). In OL-FOA, they adopt a chaotic map to initialize the population; they add the adaptive DE/best/2mutation operator to improve the exploration capability of the fruit fly approach; and finally, to improve the efficiency of the search process (by reducing the search space), the authors use the orthogonal learning mechanism. Findings To test the efficiency of the proposed approach, a test suite of 2500 web services is chosen from the public repository. To establish the competitiveness of the proposed approach, it compared against four other meta-heuristic approaches (including classical as well as state-of-the-art), namely, fruit fly optimization (FOA), differential evolution (DE), modified artificial bee colony algorithm (mABC) and global-best ABC (GABC). The empirical results show that the proposed approach outperforms its counterparts in terms of response time, latency, availability and reliability. Originality/value In this paper, the authors have developed a population-based novel approach (OL-FOA) for the QoS aware web services selection (WSS). To justify the results, the authors compared against four other meta-heuristic approaches (including classical as well as state-of-the-art), namely, fruit fly optimization (FOA), differential evolution (DE), modified artificial bee colony algorithm (mABC) and global-best ABC (GABC) over the four QoS parameter response time, latency, availability and reliability. The authors found that the approach outperforms overall competitive approaches. To satisfy all objective simultaneously, the authors would like to extend this approach in the frame of multi-objective WSS optimization problem. Further, this is declared that this paper is not submitted to any other journal or under review.

An Intelligent Web Service Composition and Resource-Optimization Method Using K-Means Clustering and Knapsack Algorithms

Service-oriented architecture (SOA) has emerged as a flexible software design style. SOA focuses on the development, use, and reuse of small, self-contained, independent blocks of code called web services that communicate over the network to perform a certain set of simple tasks. Web services are integrated as composite services to offer complex tasks and to provide the expected services and behavior in addition to fulfilling the clients’ requests according to the service-level agreement (SLA). Web service selection and composition problems have been a significant area of research to provide the expected quality of service (QoS) and to meet the clients’ expectations. This research paper presents a hybrid web service composition model to solve web service selection and composition problems and to optimize web services’ resource utilization using k-means clustering and knapsack algorithms. The proposed model aims to maximize the service compositions’ QoS and minimize the number of web services integrated within the service composition using the knapsack algorithm. Additionally, this paper aims to track the service compositions’ QoS attributes by evaluating and tracking the web services’ QoS using the reward function and, accordingly, use the k-means algorithm to decide to which cluster the web service belongs. The experimental results on a real dataset show the superiority and effectiveness of the proposed algorithm in comparison with the results of the state–action–reward–state–action (SARSA) and multistage forward search (MFS) algorithms. The experimental results show that the proposed model reduces the average time of the web service selection and composition processes to 37.02 s in comparison to 47.03 s for the SARSA algorithm and 42.72 s for the MFS algorithm. Furthermore, the average of web services’ resource utilization results increased by 4.68% using the proposed model in comparison to the resource utilization by the SARSA and MFS algorithms. In addition, the experimental results showed that the average number of service compositions using the proposed model improved by 26.04% compared with the SARSA and MFS algorithms.

QoS-based data analytic service selection: a comparative study of different learning models

QoS-based web service selection has been studied in the service computing community for some time; however, data characteristics are not considered. In this work, we have studied the use of different machine learning algorithms as meta-learners in predicting the performance of data analytic services for the given dataset. We used a meta-learning algorithm to incorporate meta-features in the selection process and we used clustering services as an example of data analytic services. We have also investigated the impact of the number of data features on the performance of the meta-learners. We found that, out of the 5 classification models, SVM showed the best results in predicting the recommended service for the given dataset with an accuracy of 78%. When it comes to regression models, MLP was the best regressor. We recommend considering only simple meta-features that can be collected for most datasets, as those proved to be sufficient to achieve good prediction accuracy.

Service selection in a marketplace: a multi-perspective solution

Most of the current research work on web service selection only considered the selection problem from the perspective of one party – service consumers. A service marketplace serves many parties including service consumers and providers. Thus, it is important to consider multiple parties. In this thesis, we propose a service selection model considering the benefits of multiple parties: consumers, providers and the marketplace. The model ranks services based on not only how much these services satisfy the user requirements but also how much the requests can be distributed to different providers and the revenue gain in the marketplace. We design different objective functions, then combine into a QoS-Plus-PF objective function. The results show that proposed model could achieve a high degree of satisfaction of user requests (i.e., 0.61% to 5.26% worse than the optimal score), and meanwhile have the capability of promoting more diversified set of services (i.e., 48.95% promotion percentage).

QoS-based data analytic service selection: a comparative study of different learning models

QoS-based web service selection has been studied in the service computing community for some time; however, data characteristics are not considered. In this work, we have studied the use of different machine learning algorithms as meta-learners in predicting the performance of data analytic services for the given dataset. We used a meta-learning algorithm to incorporate meta-features in the selection process and we used clustering services as an example of data analytic services. We have also investigated the impact of the number of data features on the performance of the meta-learners. We found that, out of the 5 classification models, SVM showed the best results in predicting the recommended service for the given dataset with an accuracy of 78%. When it comes to regression models, MLP was the best regressor. We recommend considering only simple meta-features that can be collected for most datasets, as those proved to be sufficient to achieve good prediction accuracy.

Service selection in a marketplace: a multi-perspective solution

Most of the current research work on web service selection only considered the selection problem from the perspective of one party – service consumers. A service marketplace serves many parties including service consumers and providers. Thus, it is important to consider multiple parties. In this thesis, we propose a service selection model considering the benefits of multiple parties: consumers, providers and the marketplace. The model ranks services based on not only how much these services satisfy the user requirements but also how much the requests can be distributed to different providers and the revenue gain in the marketplace. We design different objective functions, then combine into a QoS-Plus-PF objective function. The results show that proposed model could achieve a high degree of satisfaction of user requests (i.e., 0.61% to 5.26% worse than the optimal score), and meanwhile have the capability of promoting more diversified set of services (i.e., 48.95% promotion percentage).

Interactive QOS browsing for web service selection

In most of the current research works on Quality of Service (QoS) based web service selection, searching is usually the dominant way to find the desired services. This approach comes with the potential problem of framing search queries properly due to requestor's lack of knowledge or vague requirement about QoS attribute values. In this thesis, we propose an interactive QoS browsing mechanism that uses the concept of clustering to present the QoS value distribution to requestors followed by finer views of service quality. By analyzing various QoS attributes, we believe that the symbolic interval data is a proper type of representation, compared with the single valued numerical data. Therefore, we use interval data clustering algorithms to implement our browsing system. We conducted experiments on simulated QoS datasets to compare the performance of using different distance measures and show the effectiveness of the interval data clustering algorithm used. The result of the experiments show that the proposed approach provides an effective, user guided QoS based service selection approach that can conceivably overcome the problems with current approaches.