Algorithmic research is an established knowledge engineering process that has allowed researchers to identify new or significant problems, to better understand existing approaches and experimental results, and to obtain new, effective and efficient solutions. While algorithmic researchers regularly contribute to this knowledge base by proposing new problems and novel solutions, the processes currently used to share this knowledge are inefficient, resulting in unproductive overhead. Most of these publication-centred processes lack explicit high-level knowledge structures to support efficient knowledge management. The authors describe a problem-centred collaborative knowledge management architecture associated with Computational Problem Solving (CPS). Specifically we articulate the structure and flow of such knowledge by making in-depth analysis of the needs of algorithmic researchers, and then extract the ontology. We also propose a knowledge flow measurement methodology to provide human-centred evaluations of research activities within the knowledge structure. This measurement enables us to highlight active research topics and to identify influential researchers. The collaborative knowledge management architecture was realized by implementing an Open Computational Problem Solving (OpenCPS) Knowledge Portal, which is an open-source project accessible at http://www.opencps.org.

Parallel to the research e#ort spend on the investigation of reusable components for building knowledge based systems amongst which problem-solving methods playacentral role, researchers from the uncertain reasoning communityhave made signi#cant progress in the developmentofpowerful and e#- cient algorithms for uncertain reasoning based upon numerical representations of uncertainty. We argue that this important class of algorithms has been totally neglected in the research on problem-solving methods carried out during the last years. Almost all existing approaches are dominated by the examination of methods based on the paradigm of search. In this paper we try to close this gap by a systematic analysis of numerical methods for uncertain reasoning. We base our investigation on well known notions used to describe reusable problem-solving methods and show that there are many analogies that could be exploited for the development of reusable methods for uncertain reasoning. Our res...