Parameter Tuning for ABC-Based Service Composition with End-to-End QoS Constraints

Abstract

QoS-aware service composition problem has been drawn great attentions in recent years. As an NP-hard problem, high time complexity is inevitable if global optimization algorithms (such as integer programming) were adopted. Researchers applied various evolutionary algorithms to decrease the time complexity by looking for near optimum solution. However, each evolutionary algorithm has two or more parameters the value of which is to be assigned by algorithm designers and likely has impacts on the optimization results (primarily time complexity and optimality). Our experiments show that there are some dependencies between the features of service composition problems, the value of the evolutionary algorithm's parameters, and the optimization results. In this paper, we use a popular evolutionary algorithm Artificial Bee Colony (ABC) to solve service composition problem and focus on the ABC's parameter turning issue. The objective is to identify the potential dependency to help service composition algorithm designers easily set up the values of ABC parameters to obtain preferable composition solution without many times of tedious attempts. Five features of service composition problem, three ABC parameters and two metrics of the final solution are identified. Based on a large volume of experiment data, ABC parameter tuning for a given service composition problem is conducted using C4.5 algorithm and the dependency between problem features and ABC parameters are established using multiple linear regression method. An experiment on a validation dataset shows the feasibility of our approach.