S-ABC: Domain-oriented Artificial Bee Colony Algorithm For Service Composition And Its Support Tool

With the rapid development and maturity of service computing, cloud computing, Internet of things and other advanced technology, large scale intelligent services have emerged in various fields. And the user requirements are becoming more and more complex, a single service has been unable to meet them, which requires service composition according to the user demands and eventually obtains the composite service to meet the needs of users. So, it is becoming a hotspot of the research that how to build the service composition solution from these large scale intelligent services based on the user requirements, which has caused wide attention. At the same time, with the research on big data being carried on, different domains show some features. Many service domains have their own features and operation rules, and these features play an important role in the service composit ion problem. Therefore, it becomes very important that how to utilize the domain features to improve the efficiency and effects of a service composition problem in the new service environment. In view of the two problems above, the domain-oriented artificial bee colony algorithm S-ABC for service compositions is proposed in this paper, and a detail research has been carried out from the following three aspects:(1)The superiorit y analysis of the artificial bee colony algorithm for service composition problem: in order to solve the service composition problem preferably by adopting the artificial bee colony algorithm, and we analyze the algorithm in detail, including the divisions of candidate service spaces, the food source generation, the neighborhood search, the algorithm arbitration rules and so on.(2)The research on the basic theories of the S-ABCsc paradigm and the algorithms: in order to study the effects of service domain features have on the efficiency of service composition problems, three service domain features are proposed, including the Priori, Similarity and Correlation. Based on the three domain features, we analyze the service composition problems and their solving algorithms, thus form the domain-oriented artificial bee colony algorithm S-ABCsc paradigm for service composit ion through combing these three domain features and the artificial bee colony algorithm. The paradigm puts forwards the artificial bee colony algorithm framework for service composit ions, including the search strategy of service spaces, the food source generation strategy for service compositions, the fitness function for service composit ions, the employed bees phase for service compositions, the observed bees phase for service composit ions, the scout bees phase for service compositions, and the algorithm arbitration rules. Among them, according to the two kinds of search strategies of service spaces, two types of algorithms are formed, that is, the S-ABCsc algorithm based on the priority search strategy and the S-ABCsc algorithm based on the balanced search strategy. In addition, in order to analyze the use conditions of these two kinds of search strategies of service spaces, some specific metrics are proposed for Priori and Similarity, and a large number of experiments are conducted to summarize the related rules.(3)The design and development of the support tool of the S-ABCsc paradiam, in order to adopt the S-ABCsc paradigm to solve the service composition problem more conveniently, we develop the support tools of the paradigm. The modules of the system include algorithm prefabrication and execution, service process management, candidate service management, requirement classificatio n management and historical data management.