Fault-tolerant execution planning for collaborative business processes based on genetic algorithms

In the present study, we developed a method that provides, while minimizing costs, guaranteed-reliable execution plans for collaborative business processes conducted, via, web services. To that end, physical- and time-redundancy techniques are utilized and dynamic modifications of execution plan are provided. In order to address the dynamic execution planning problem, known to be NP-hard, we also developed a Genetic Algorithm (GA), the effectiveness of which was demonstrated through a set of experiments. Specifically, the GA was shown to be capable of providing near-optimal solutions in polynomial time. The main contribution of this paper is the more general execution planning method developed in the present study. While previous research assumed that the execution cost, time, and reliability of web services are the same, we relaxed that assumption. We expect that this will facilitate the application of our method in practice.