The Configurability Study on Artificial Bee Colony Algorithm for QoS-aware Service Composition

Artificial bee colony algorithm (ABC) has been widely adopted in solving QoS-aware service composition problem due to its advantages of few parameters, strong robustness and search capability. However, how to configure the parameters and optimization strategies of ABC to get the best optimization result is a challenge for non-professional users. To attach this issue, this paper focuses on the configurability study of artificial bee colony algorithm and extracting seven configurable points, namely key parameters assignments, initial food sources generation strategies, neighborhood search strategies, fitness function, food source selection methods for onlooker bees, new food source generation methods for scout bees and stop conditions, meanwhile, several different optional strategies for each configurable point are summarized. A prototype system for configuring ABC points is realized, users can obtain a concrete algorithm for QoS-aware service composition by determining specific strategy for each configuration point according to their application demands. In addition, the results and the performance analysis of the produced algorithm can be viewed.