Concurrent Engineering

~ ~ There are several research efforts into computer support for CE currently underway. The best known of these is the DARPA Initiative in Concurrent Engineering (DICE) at the Concurrent Engineering Research Center (CERC) at West Virginia University. Other research efforts include the CAD Framework Initiative (CFI), the Open Systems Architecture for Computer-Integrated Manufacturing (CIM-OSA), and the Engineering Information System (EIS). These four research efforts are reviewed in [Jagannathan et al 1991]

Jagannathan et al identify three areas in which CE research is progressing: Management Processes, Technical Practices, and Information Technology. They then focus on the Information Technology issue, which the four previously mentioned systems support. Within this category, they identify five areas in which computers can be used to support CE: sharing information, collocating people and programs, integrating tools and services with frameworks, coordinating the team, and capturing corporate history.

Sharing information is necessary to promote cooperation among the members of multi-disciplinary design teams. The models each of the members uses in the design process may be quite different, however, and the system should be able to translate models produced by one designer into the format and vocabulary used by another.

Collocating people and programs is achievable by networking. The key to maintaining this virtual collocation is in making the access to programs, people, and data across the network transparent to the user.

Integrating tools and services with frameworks is a means of allowing designers to use different tools with ease. The idea is to support a single means of user interaction with all of the tools, to make them more uniform. This would lead to less training time on new tools and a greater chance that designers would use all of the tools available to them.

Coordinating the team is concerned with keeping all members of the design team apprised of the current state of the design. Whenever a design decision is proposed, all members of the team are informed.

Capturing corporate history is a means of keeping track of design decisions and the reasons for them. It uses an electronic design notebook, or some other means of recording decisions.
Londoño, et al [1989] are doing research to build a system to support CE for DICE. This system is designed to help designers participate in cooperative design. They have chosen to use a blackboard for communicating and for control of information flow. All data about parts is held in the Product, Process and Organization (PPO) database. This database is accessible by all of the product's designers. Individual designers use Local Object Workspaces (LOWs), in which they can modify current designs and test hypotheses before committing everyone on the design team to a decision. The whole design process is overseen by the Project Lead (PL). The PL must be able to keep track of new tasks and follow the progress of the tasks, as well as generate new tasks, in order to change the focus of the design, and assign them to designers.

Representations of the designed products are stored in frames and pointers to AutoCAD files. This information is stored in the PPO database. Designers must have access to all the information in the PPO database. The database should support user friendly searches for parts on which a designer wishes to work. Though the system defaults to helping the designer in a bottom-up design fashion, there exists the option to do all or part of the design top-down. Initialization of the blackboard includes adding design specifications, determining dependencies, deciding on initial tasks, and including known heuristics.

Global, local, and implicit constraints are supported by the system. Global constraints refer to overall specifications that apply to the design as a whole. Local constraints refer only to individual parts. Implicit constraints are constraints which need to be inferred from the constraints on subparts. Reasoning with constraints can be done to allow the system to spawn new tasks to be done. Dependencies are kept by the system to indicate the other aspects of the design that may be affected by a design decision about a part. Those on the dependency list are notified when a change occurs. Maintenance of versions has not yet been addressed in the system.

Heuristics are another form of design knowledge used in the system. They can be used to help schedule tasks and develop uniform plans of action during a design. Negotiation is an important consideration involving multiple designers. The designers are allowed to vote on accepting a change to the design.
The Function Advisor (see [Kott et al 1991]) is another CE support system being developed. Their main concern is with allowing ``a computer-based advisory system to support the cooperation between multiple engineering agents.'' The research issues include: planning and management of the activities of multiple agents; representing and modeling the product; managing multiple representations and versions of the product; developing Engineering Databases capable of supporting Concurrent Engineering; managing and propagating constraints and avoiding inconsistencies in the current state of the product description; sharing the information between the multiple agents without creating an ``information blizzard.''

The objective of the system they are building is to help a group of knowledge workers (designers) design an engine and all of its component parts. They categorize the functional objectives of the Function Advisor they are developing into four groups: advising the human organizer of tasks to be performed; retrieving, organizing, and conveying pre-stored information relevant to the current needs of the organizer; inferring information that is not explicitly stored; and monitoring the design for consistency, completeness, and correctness.

~ ~ The more specific functional goals for the Function Advisor are given below:

1. Provide an inference mechanism for reasoning about the design organization.
2. Alert human workers of abnormalities or inconsistencies between the design and the design goals.
3. Suggest a plan of design activities to human designers.
4. Advise the human designer on dependencies between parts of the product.
5. Find parts, constraints, or other design knowledge of use to a particular designer.
6. Prevent attempts to finalize a design before designs which affect it are finalized.
7. Collect and store all product documentation.
8. Detect those parts which are dependent on other design decisions.
9. Identify the need for special purpose design aids.
10. Guard against re-occurring design errors.

Subramanian, et al [1990] are developing a design support environment for CE. They used the system as a testbed for their theories about which kinds of information are used during group design. When different groups work on a design from different aspects the design activity is usually done concurrently. They propose that a computer system designed to help groups from the various aspects of product design should work similarly to the way that those groups do. The groups share computations, figures, and any other data they feel necessary. This means having mechanisms to represent designs in different ways and to maintain copies of all of the current design data. It also requires a means of scheduling the use of the data.
Knowledge-based systems which communicate by means of a blackboard are used in a project called DICE (Distributed and Integrated environment for Computer-aided Engineering) (see [Sriram et al 1989]). This project is not associated with the DARPA Initiative in CE. Sriram's DICE is made up of a control mechanism to coordinate the activities of its modules and users, a blackboard to post the current state of the design, and knowledge modules to represent various aspects of the design. The system uses an object oriented approach to design. It allows for negotiation between agents. It provides support for the databases that are needed for this kind of distributed knowledge design. The knowledge modules have various roles: Strategy (helps the control mechanism by determining the next course of action), Specialist (an aggregation of expert systems to help make decisions), Critics (keep track of consistency in the design), and Quantitative (algorithmic evaluation or CAD tool). The Blackboard is divided into three partitions: Coordination (keeps bookkeeping information), Solution (contains the object hierarchy and current design), and Negotiation (contains constraints on the design and a trace of the negotiation process). All of this allows the user to interact with various autonomous agents when designing part of a building.
Some of the tools which could be included in a CE system are mentioned in [Boothroyd & Dewhurst 1988] and [Lemon et al 1990]. Both sets of tools are concerned with cost estimation over the entire product's life cycle during the design phase. Incorporating these tools and their methods into design systems allows designers access to knowledge about other aspects of the product's life-cycle and the processes to be used by those aspects. Since cost is affected by the processes chosen in manufacturing, assembly and disposability as well as material selection and other, more concrete costs, analysis of life-cycle costs gives designers a means of focusing on other life-cycle aspects.
The issues concerned with version control are covered in [Katz 1990]. In his work, Katz attempts to unify terms used in version management and propose a scheme for developing a complete version management system. The details are quite extensive. What he proposes is an organization in the database which connects versions of different parts together to form a design. But it also connects those versions to their parents and offspring. So the entire database will contain links to different versions of the parts, and these can be combined in different ways to obtain different current designs. These links can be used to search for alternatives and to propagate constraints to all versions of a part.
There are several different thrusts in the area of computer support for CE. As Jagannathan et al state, CERC and the DICE projects concentrate on the availability of information and information sharing. Their main goal is to provide as many people as possible with as much useful data as possible so as to make decision-making easier. Jagannathan et al state that this is also the thrust of the CAD Framework Initiative (CFI), the Open Systems Architecture for Computer-Integrated Manufacturing (CIM-OSA), and the Engineering Information System (EIS). Londoño et al are also working to support this view. They are also concerned with supporting negotiation.

Subramanian et al are also concerned with sharing information among the members of the design team and supporting this kind of interchange. Sriram et al are concerned with communication between intelligent agents involved in the design. They are also concerned with the computer's handling of negotiation in the event that some decision is not amenable to all parties. Kott et al concentrate on using the computer to keep track of the design, maintain the focus, and monitor the consistency of the design.

This section has attempted to present a summary of some of the current CE research, and has tried to reflect what their main concerns are. It has also attempted to characterize (perhaps wrongly) what they consider as the key aspects to the problem of producing CE systems.