Research On Knowledge Management Technologies For Software Process Improvement

The effective management of software process assets is the key foundation for Software Process Improvement (SPI) and sustainable development of software organizations. Nowadays, Knowledge Management (KM) is attracting extensive attention in domains of Management, Artifical Intelligence and Software Engineering. One hot research topic is the integration of knowledge management process with business processes for improving the distribution and reuse of knowledge assets within organizations.SPI is an important kind of practices for sustaining and improving software engineering capabilities, which relates to most of the main business activities in software organizations. Aiming at providing supports for SPI, this dissertation carries out systematic research and practice on KM technologies. The main contributions of this dissertation include:(1) A Knowledge Management framework for Software Process Improement. The key features of KM in knowledge intensive organizations are identified. Then based on the CMM model, software processes, process assets and the relations between them are discussed. A SPI system is studied based on which software organization's business activities are discussed. Organizational-level and project-level knowledge management requirements are identified respectively. Then a KM framework for SPI is proposed. This framework emphasizes the idea of connecting knowledge services and business activities, and it provides an important foundation for the research on knowledge management technologies.(2) An innovative knowledge flow (KF) modeling and control approach. An extention to WfMC (Workflow Management Coalition)'s workflow process definition meta model is presented. Based on that, modeling and control of knowledge flows are studied. A knowledge flow modeling approach is proposed, with five kinds of knowledge flow components to represent knowledge distribution and reuse, communication and cooperation among participants. To deal with dynamic elements in knowledge flows, an adaptive knowledge flow control approach withcorresponding algorithms are developed based on resource constraints, changes of knowledge requirements, and time constraints. This works helps to implememt KM in business process, and helps to achieve the aims of active and durative knowledge distribution.(3) Mining and retrieval approaches on software process assets. Mining and retrieval of software process assets act as the basis of decision supports for SPI related activies. The representation of software process assets is studied first, and an organizational practice framework is designed to represent the relations among the elements of knowledge assets. Then, several practical process assets mining tasks and approaches are developed. With the mined rules and actor's profile, the retrieval approaches of software process assets are designed.These approaches help to improve the degree of knowledge assets reuse.(4) A knowledge management engine for SPI. This engine is the implementation of the core components of the above KM Framework for SPI. The design of the engine integrates technologies including knowledge flow, OLAP, data mining and information retrieval. The engine is integrated with a CMM-Based Software Quality Assurance Platform, providing knowledge services for organizational- and project-level activies of software processes. It improves the accumulation, distribution and reuse of process assets, and provides support to the realization of SPI's aim. By practice, the idea and technologies of the KM engine's design gets positive response.