Abstract. To direct automated Web service composition, it is compelling to provide a template, workflow or scaffolding that dictates the ways in which services can be composed. In this paper we present an approach to Web service composition that builds on work using AI planning, and more specifically Hierarchical Task Networks (HTNs), for Web service composition. A significant advantage of our approach is that it provides much of the how-to knowledge of a choreography while enabling customization and optimization of integrated Web service selection and composition based upon the needs of the specific problem, the preferences of the customer, and the available services. Many customers must also be concerned with enforcement of regulations, perhaps in the form of corporate policies and/or government regulations. Regulations are traditionally enforced at design time by verifying that a workflow or composition adheres to regulations. Our approach supports customization, optimization and regulation enforcement all at composition construction time. To maximize efficiency, we have developed novel search heuristics together with a branch and bound search algorithm that enable the generation of high quality compositions with the performance of state-of-the-art planning systems. 1

We present Shopper, a plan execution engine that facilitates experimental evaluation of plans and makes it easier for planning researchers to incorporate replanning. ShopperinterpretstheLTMLplanlanguage,whichextendsPDDL in two major ways: with more expressive control structures,andwithsupportforsemanticwebservicesmodeledon OWL-S. LTML’s command structures include not only conventional ones such as branching, iteration, and procedure calls, but also features needed to handle HTN plans, such as precondition-filtered method choice. Unlike conventional programminglanguages,LTMLsupportsinteractionwiththe agent’s belief store, so that its execution semantics line up withthoseassumedbyplanners. LTMLactionsextendPDDL actionsinhavingoutputsaswellaseffects,whichmeansthat they can support actions that sense the world; an important special case of this is semantic web services, which reveal information about a state hidden from the agent. To support experimentation as well as action in the real world, Shopper accommodates multiple, swappable implementations of its primitive action API. For example, one may interact with real web services through SOAP and WSDL, or with simulated web services through local procedure calls. We describe novel features of LTML, the interpretation strategy, swappable back-ends, and theimplementation.

This paper presents a distributed architecture for automating data mining processes using standard languages. Data mining is a difficult task that relies on an exploratory and analytic process of processing large quantities of data in order to discover meaningful patterns. The increasing heterogeneity and complexity of available data requires some expert knowledge on how to combine the multiple and alternative data mining tasks to process the data. Here, we describe data-mining tasks in terms of Automated Planning, which allows us to automate the data-mining knowledge flow construction. The work is based on the use of standards that have been defined in both data mining and automated-planning communities. Thus, we use PMML (Predictive Model Markup Language) to describe data mining tasks. From the PMML, a problem description in PDDL (Planning Domain Definition Language) can be generated, so any current planning system can be used to generate a plan. This plan is, again, translated to a data-mining workflow description, KFML format (Knowledge Flow file for the WEKA tool), so the plan or data-mining workflow can be executed in WEKA (Waikato Environment for Knowledge Analysis).

... of services, typically in a SOA, with a precise mathematical meaning in a formal ontology. These annotations allow a higher degree of automation. The last decade has seen a wide proliferation of such approaches, proposing different ontology languages, and paradigms for employing these in practice. The next chapter gives an overview of these approaches. In the present chapter, we provide an understanding of the fundamental techniques, from Artificial Intelligence and Databases, on which they are built. We give a concise, ontology-language independent, overview of the techniques most frequently used to automate service discovery and composition.