A Memetic Algorithm with Distance-guided Crossover: Distributed Data-intensive Web Service Composition

Data-intensive Web service composition (DWSC) combines numerous interoperating, distributed and reusable data-intensive services to accomplish a specific task. In this paper, we will propose a Memetic Algorithm (MA) with a novel crossover operator to solve the distributed DWSC problem. In particular, we define two heuristics which enable explicit use of two important sources of information in producing offspring solutions in the MA. 1) location information from the problem to quickly determine the suitability of any solution and 2) useful information embedded in sub-solutions (e.g. common parts of the solutions). Experiments have been performed on benchmark datasets and the results show that our method is more effective than a state-of-the-art baseline method.