Dynamic Asynchronous Aggregate Search for Solving QoS Compositions of Web Services

There has been recently an increase of interests in distributed constraint satisfaction (DisCSP) algorithms. Most of these algorithms assume that the set of variables and constraints in a DisCSP are completely known and fixed. However, these assumptions do not hold in many open environments where unexpected events and changes can happen. Therefore this issue needs to be addressed in order for a DisCSP algorithm to be practically used in real life applications. Our contribution in this paper is two-fold. Firstly, we present an extended version of the asynchronous aggregate search (AAS), called the dynamic asynchronous aggregate search (DynAAS) - as a new algorithm for dynamic and uncertain environments. Secondly, we investigate the applicability of DynAAS in the Web service domain and argue that DynAAS is suitable for solving the QoS compositions of Web services. We also carry out the experiments to show the efficiency of our proposed algorithm in solving the QoS composition problem