The smart workflow foundation

This report presents the Smart Workflow Foundation (SWF), a new architecture which adds constraint solving capabilities to workflow engines. Thanks to that extension, workflow definitions are freed from low level implementation details and can benefit from smart and robust resource allocation. This architecture represents a radical change over classical engines where the execution of each task or procedural step is either pre-assigned to some entity, e.g.,employee, either computed by a former task. The proposed system uses abstract workflow definitions combined with a characterizing of the resources to efficiently match tasks requirements to resources abilities and availabilities. During this process, the possible future steps of a workflow are considered. This mimics the capabilities of human beings, able to infer the consequences of a decision against some foreseeable future. The system is built on top of Windows Workflow Foundation and evaluated through several simulations. Various extensions are also presented in order to improve the scope of the reasoning, automatically drive the execution flow from the result of high level optimization problems, and use the newly proposed abstraction to solve capacity planning scenarios.