A precise definition is provided for general configuration tasks. Two important assumptions are identified: (i) functional architecture and (ii) key component per function. A domain-independent model is presented based on these assumptions. These assumptions are shown to be both useful and tenable in real domains. They are useful because they limit the complexity of the general configuration task, determine the basic knowledge needed for solving a configuration task, and enable more efficient problem solving methods. Ideas are presented both for representing the knowledge and controlling the search. Some of these ideas were originally implemented in the Cossack expert system.

/pdf/towards-a-generic-model-of-configuraton-tasks-vaqj83nsxo.pdf

Towards a generic model of configuraton tasks

Our experience with implementing the LAN Designer system for cooperative group design of local area networks has led to a number of insights into how cooperative group design systems can be structured to effectively support this process. Support for conflict management appears to be critical for real-world cooperative design; general conflict detection and resolution heuristics can effectively support this; and future progress in supporting cooperative design requires design models that support self- understanding (i.e. rich design rationale) and self-modifiability (i.e. a powerful replanning facility) as well as the ability to analyze and critique indefinitely described designs. The LAN Designer system is an instantiation of a generic cooperative group design support shell we have developed. We are currently extending this shell to support the interaction of both human and machine-based design agents, and to learn from the process of doing so. We plan to instantiate the shell in a number of different cooperative design and planning domains in order to evaluate and enhance its effectiveness and breadth of applicability.

Insight into Cooperative Group Design: Experience with the LAN Designer System

RI and beyond: AI technology transfer at DEC

The research presented here aims at providing a comprehensive framework for solving configuration problems, based on the Constraint Satisfaction paradigm. This thesis is addressing the two main issues raised by a configuration task: modeling the problem and solving it efficiently. Our approach subsumes previous approaches, incorporating both Simplification and further extension, offering increased representational power and efficiency. 
Modeling. We advance the idea of local, context independent models for the types of objects in the application domain, and show how the model of an artifact can be built as a composition of local models of the constituent parts. Our modeling technique integrates two mechanisms for dealing with complexity, namely composition and abstraction. Using concepts such as locality, aggregation and inheritance, it offers support and guidance as to the appropriate content and organization of the domain knowledge, thus making knowledge specification and representation less error prone, and knowledge maintenance much easier. 
There are two specific aspects which make modeling configuration problems challenging: the complexity and heterogeneity of relations that must be expressed, manipulated and maintained, and the dynamic nature of the configuration process. We address these issues by introducing Composite Constraint Satisfaction Problems, a new, nonstandard class of problems which extends the classic Constraint Satisfaction paradigm. 
Efficiency. For the purpose of the work presented here, we are only interested in providing a guaranteed optimal solution to a configuration problem. To achieve this goal, our research focused on two complementary directions. 
The first one led to a powerful search algorithm called Maintaining Arc Consistency Extended (MACE). By maintaining arc consistency and taking advantage of the problem structure, MACE turned out to be one of the best general purpose CSP search algorithms to date. 
The second research direction aimed at reducing the search effort involved in proving the optimality of the proposed solution by making use of information which is specific to individual configuration problems. By adding redundant specialized constraints, the algorithm improves dramatically the lower bound computation. Using abstraction through focusing only on relevant features allows the algorithm to take advantage of context-dependent interchangeability between component instances and discard equivalent solutions, involving the same cost as solutions that have already been explored.

/pdf/a-constraint-based-framework-for-configuration-2hpvg7m4vp.pdf

A constraint-based framework for configuration

MAPCon II is the second generation (Muralidhar and Irish, IEEE Journal on Selected Areas in Commumcctions 6(5), 869–873, 1988) of an expert system that interactively guides a user in performing off-line configuration for local area networks that use MAP, the manufacturing automation protocol. This paper describes the configuration task in general and MAPCon in particular. Though MAPCon's purpose is off-line configuration, its problem domain requires that it accomplish other reasoning objectives in addition to those commonly associated with configuration. It is in the process of being expanded into an on-line network supervisor. We develop a taxonomy of reasoning objectives and show how MAPCon combines two different kinds of reasoning to accomplish its objectives. Our experience confirms that of other researchers, and suggests that building robust, practical systems will require us to understand more clearly the interfaces among different reasoning objectives. The paper has four parts: 1. a definition of configuration and other reasoning objectives; 2. a summary of the problem domain in which MAPCon operates; 3. a description of MAPCon as the user sees it; 4. a look ‘under the hood’ to see how MAPCon combines different objectives.

K. H. Muralidhar

Papers

MAPCon: a case study in a configuration expert system