An inherent difficulty in enumerative search algorithms for optimisation is the combinatorial explosion that occurs when increasing the size of the input. Among incomplete algorithms that address this issue, ant colony optimization(ACO) uses a combination of random and heuristic methods plus reinforcement learning, which proved efficient on a wide range of CSPs problems. This paper presents results in applying an ACO-based algorithm to configuration, which to the best of our knowledge was never investigated before. We describe how the nature of unbounded configuration problems impacts the ACO approach due to the presence of set-variables with open domains. We propose an ACO framework able to deal with those issues through an original pheromone model and algorithm. We also present the use of particle swarm optimization (PSO) to converge towards good parameter sets. Finally, we provide early experimental results, both for random problem instances andthe "racks" optimisation problem.

Semantic Web Service (SWS) composition is a challenging AI problem. We describe a theoretical and experimental framework based upon finite model search for constrained object models to address this problem. In many AI situations the input is rather simple, and the results complex to obtain. SWS composition requests themselves can turn very complex, and the problem of building these requests can be viewed as an AI problem of its own. This paper presents an operational end to end approach to composing/publishing Semantic Web Services involving two main reasoning stages. Composing is first performed at the abstract level of goals (each roughly representing a discovery request), which yields a composition request at the workflow level. The resulting workflow is finally processed to generate a valid publishable semantic web service. We present experimental results obtained on industrial use cases during the DIP project.

Automatic or assisted workflow composition is a field of intense research for applications to the World Wide Web or to business process modeling. Workflow composition is traditionally addressed in various ways, generally via theorem proving techniques. The originality of this research stems from the observation that building a composite workflow bears strong relationships with finite model search, and that some workflow languages can be defined as constrained object metamodels (van der Aalst et al., 2003). This leads to consider the viability of applying configuration techniques to this problem. Our main contribution is to prove the feasibility of such an approach, with some advantages and drawbacks compared to logical based techniques. We present a constrained object model for workflow composition, based upon a metamodel for workflows and ontologies for processes and data flows. Experimental results are listed for a working implementation that generates complex interleaving composite workflows involving transformations, synchronization and branching constructs.