Supporting the Modelling Life Cycle Through Knowledge Management

Models have been widely used by scientists to address environmental problems. Models are representations of reality whose goal is the abstraction of irrelevant details of the phenomena or processes being modelled. In order to create an environmental model, it is imperative a modeling process that establishes a model life cycle in which as intellectual creativity as abstraction skills are required. Nowadays, environmental models have been the target of research once they help organizing ideas, understanding data, communicating and testing hypotheses as well as accomplishing predictions. In spite of some environmental models are available in the Internet, the knowledge acquired in the modeling process is difficult to be found or even it does not exist. That knowledge is important for the scientist to determine if the model is adequate or not to address his/her problem. Environmental models, as data, need to be handled as scientific organizations resources. In order to support this we propose MODENA (Model Development Environment), a web tool for scientific models management and execution using Computational Grid platform. This environment is composed of two systems: ModManager and ModRunner. ModManager deals with knowledge management about scientific models, acting as a scientific model library that allows cataloguing, searching, reutilization and generation of new models. ModRunner deals with the management of the execution of scientific models in a Grid environment, allowing model composition to generate a scientific Grid Workflow to be executed by distributed services offered by Grid Services. In this work we propose a guideline to support the first two steps of the modelling life cycle (to support the modelling and the management of the knowledge about models): problem identification and model creation. This guideline is intended to help either the researcher to choose the model that best fits the requirements or the modeler to develop his/her own model. All the steps are supported by the MODENA environment, which provides support for the modelling life cycle and the management of the knowledge obtained in it. After problem identification, the researcher may track a guideline to help model creation. Every kind of information and knowledge about a model being created or identified is captured by the guideline and registered in the system. The aim is to identify every component of the model in order to do the best modelling, and provide sufficient information to model use and feedback. Among the components, we could cite model parameters, variables, constants, equations, constraints, “business rules”, implemented programs and so on. The use of knowledge management can bring new possibilities to model identification, registration, utilization, reuse, learning, and sharing. Furthermore, the use of the proposed guidelines may lead to better model description and reuse, increasing modelling productivity and efficiency. As future work we envision the use of model ontologies to improve model search and composition, once the semantics of model description could improve the quality of search results and facilitate model generation through new possibilities of model composition. We also intend to integrate MODENA with GCC (Scientific Knowledge Management), a software that provides many kinds of facilities to knowledge management in scientific work.