A Framework for Mutant Genetic Generation for WS-BPEL Juan-Jose Dominguez-Jimenez, Antonia Estero-Botaro, and Inmaculada Medina-Bulo University of Cadiz C/ Chile 1, 11003, Cadiz, Spain {juanjose.dominguez,antonia.estero,inmaculada.medina}@uca.es Abstract. The rise of Web Services and their WS-BPEL compositions in recent years makes necessary to pay special attention to testing in this context. Muta- tion testing is a white box testing technique that has been applied successfully to programs written in different languages. In order to apply it we need a set of mutation operators and a system for mutant generation. This paper introduces a set of mutation operators for the WS-BPEL 2.0 language and a framework, based in genetic algorithms, for automatic mutant generation without rendering all pos- sible mutants. This framework can also detect potentially equivalent mutants. 1 Introduction Advances in testing techniques are intimately linked to emerging trends in software development. One of the latest trends is marked by the emergence of so-called web ser- vices (WS). These allow rapid application development, characterized by a low cost and an easy distributed application composing [1]. The OASIS WS-BPEL 2.0 standard [2] has become the industry reference for WS compositions. This allows you to specify the logic of the service composition (messaging, synchronization, iteration, treatment of erroneous transactions, etc.) regardless of service implementation. White box testing techniques, and more specifically mutation testing [3, 4], depend on the programming language used in software development. This technique can and should play an important role in defining test strategies for WS compositions. In this sense, it is necessary to develop a system that can generate mutants for WS-BPEL. For mutant generation we need a mutation operators set. To the best of our knowl- edge there is no work dealing with the definition of mutation operators for the WS- BPEL language. Moreover, the specific syntactical and semantical features of this lan- guage make necessary to define its own operators. One of the main drawbacks of mutation testing is the high computational cost in- volved. This is due to we usually generate a large number of mutants. Therefore, we think that can be interesting the application of optimization techniques, such as genetic algorithms (GA) [5], in mutant generation. This paper presents a set of mutation operators and a framework for automatic mu- tant generation for WS-BPEL service compositions using a GA. Our work shows a novel use of GAs when applied to mutant generation. The algorithm also detects po- tentially equivalent mutants. The structure of the paper is as follows: Section 2 briefly