A comparison of modeling approaches for planning in Cyber Physical Production Systems
One of the largest obstacles for the automation of web services, as a composition of services that allows for the decrease in efforts, time, and cost in their manual development, integration, and maintenance, is finding a way to handle this planning problem, which is far from trivial. At the time of this article (2004), there were two approaches to tackle this problem. First, most of the giants of industry proposed a low level process modeling and execution language. The benefit to this approach was that it allowed programmers to create complex web services as distributed processes and to connect them in a simple way: like simple conditionals and loops. The drawback was that new processes/definition that interact/link to already defined components had to be manually programmed in, this is not only challenging but time consuming and error prone. Second, research from within the Semantic Web community proposed a top-down unambiguous description of web services capabilities. The benefit of this approach is that it enables the use of reasoning in web services, and to automate web service tasks such as discovery and composition. They major challenge is the ability to automatically compose services that can directly execute, if this was achieved, it would greatly reduce effort, time, and errors dues to manual composition.