Vadim Ermolayev

An agent-oriented model of a dynamic engineering design process

One way to make engineering design effective and efficient is to make its processes flexible i.e. self-adjusting, self-configuring, and self-optimizing at run time. This paper presents the descriptive part of the Dynamic Engineering Design Process (DEDP) modelling framework developed in the PSI 1 project. The project aims to build a software tool to assist managers to analyse and enhance the productivity of the DEDPs through process simulations. The framework incorporates the models of teams and actors, tasks and activities as well as design artefacts as the major interrelated parts. DEDPs are modelled as weakly defined flows of tasks and atomic activities that may only become apparent at run time because of several presented dynamic factors. The processes are self-formed through the mechanisms of collaboration in the dynamic team of actors. These mechanisms are based on contracting negotiations. DEDP productivity is assessed by the Units of Welfare collected by the multi-agent system that models the design team. The models of the framework are formalized in the family of PSI ontologies.

Evaluation of Semi-automated Ontology Instance Migration

Ontology instance migration is one of the challenges in knowledge management. It becomes even more complex in distributed settings when, for example, several autonomous agents use partial assertional knowledge in a domain that is formalized by different though semantically overlapping descriptive theories. Agents exchange instances of their ontologies when cooperate. Such an exchange is essentially the migration of the assertional part of an ontology to other ontologies owned by different agents. The paper presents our method and tool support for migrating instances between different semantically overlapping ontologies. The method is based on the use of manually coded formal rules describing the changes between the input and the output ontologies. The tool to support the process is implemented as a plug-in to Cadence ProjectNavigator software. The main contribution of the paper is in presenting the results of the evaluation of this tool. It reports about the setup for our evaluation experiments, the metrics used for measuring the quality of instance migration, the ontologies that have been chosen as the experimental data, and the evaluation results. Evaluation results are satisfactory and suggest some directions for the future work.

Increasing Predictability and Sharing Tacit Knowledge in Electronic Design

The chapter reports on the use of knowledge process learning, articulation and sharing technologies developed in the ACTIVE project for increasing the performance and decreasing the ramp-up efforts of knowledge workers in engineering design projects. Attention is paid to the specific characteristics of knowledge processes in microelectronic engineering design, of which one of the most important is the absence of predefined workflows. Instead of following rigid working patterns, the knowledge workers exploit their tacit knowledge and experience for finding the most productive way through the “terrain” of the possible process continuations. The knowledge workers in this domain are design project managers, designers, and design support engineers. Process knowledge is mined from distributed heterogeneous datasets, fused, and used for visualizing the design project plan and execution information. The visualization suggests optimized performance, points to the bottlenecks in executions, and fosters collaboration in development teams. A project navigation paradigm is developed that helps knowledge workers more easily accomplish their work. We describe the software prototype architecture and implementation. Validation results are presented which indicate that the solution is helpful in providing expert assistance to design project managers performing their typical tasks of project planning and execution control.

Articulation and Sharing of Distributed Design Project and Process Knowledge

The paper reports on the work in the case study of the ACTIVE project on the use of the combination of knowledge process learning, articulation and sharing technologies for increasing the performance and decreasing the ramp-up efforts in engineering design projects. This knowledge is mined from distributed heterogeneous datasets, fused, and further used for visualizing design project plan information in a way that suggests optimized project plans and fosters collaboration on these knowledge structures in development teams. Software demonstrator architecture, implementation and validation are presented. Validation results indicate that the solution is effective in providing expert assistance to design project managers in performing their typical planning tasks.