Agent-Based Framework for Concurrent Engineering Software

In this paper, we propose a novel software interoperability framework, termed “agent-based”, that addresses many important issues critical to effective concurrent engineering. We demonstrate how concurrent engineering is supported using Designworld, an environment for total product engineering built using the agent-based framework. 1.0 Introduction Today’s software infrastructures are inadequate to support concurrent engineering effectively. Heterogeneous software tools typically do not interoperate well. As a result, many interoperability standards are beginning to emerge from the industry. CORBA (Common Object Request Broker Architecture from OMG consortium) aims at making data format independent across platforms and service location transparent to the user. BMS (Broadcast Message Server from Hewlett-Packard) provides a basic message brokering service based on message types. CFI (CAD Framework Initiative) provides common presentation and management services. However, these standards are far from sufficient to address important issues in concurrent engineering: how to share design information and knowledge between tools that operate in different disciplines with different conceptualizations; how to allow partial designs to be shared thereby reducing delay in making information available; how to enable arbitrary constraints to be disseminated, thereby avoiding premature commitments; how to avoid making assumptions (for instance, about the availability of specific services) and hardwiring them inside tools that will operate in a highly evolving engineering environment. To date, the burden of making engineering software programs interoperate and integrating their services is still largely placed on engineers and tool builders. In this paper we propose a novel framework, termed “agent-based”, that aims at facilitating information propagation and collaborative problem solving among software tools in a concurrent engineering environment, while relieving the tool user and builder of as much burden as possible. 1. Stanford University / Computer Science Department / Logic Group 2. Hewlett-Packard Labs / Software Technology Laboratory 3. Hewlett-Packard Labs / Software Technology Laboratory

[1]  V. Lifschitz,et al.  IN COLLABORATION WITH , 1992 .

[2]  Timothy W. Finin,et al.  An Overview of KQML: A Knowledge Query and Manipulation Language , 1991 .

[3]  Michael R. Genesereth Designworld , 1991, Proceedings. 1991 IEEE International Conference on Robotics and Automation.