|
|
|
|
First Generation: Rules-Based The first generation of configuration systems were built using rules-based expert systems, such as DEC’s XCON. These systems took the first leap forward by automating tedious manual processes, but were extremely difficult to maintain owing to complex conditional structures that mixed configuration logic with data. There was no notion of a product database since the data was built right into the rule structure. If a new product component was added, then someone maintaining the solution would have to find each place to insert the update into the nested rule structure. In addition, the rules based systems are not able to accommodate the modeling logic necessary for mission driven configuration and automated output generation. They have proven to be both costly and difficult to maintain and relatively unsophisticated. Second Generation: Constraint-Based Given the limitations of the rules-based first-generation systems, a new approach evolved using resource constraints, which allowed companies to separate logic from data. Second-generation systems model component data using object-oriented techniques and configure those components using global constraints. Structurally, these constraint-based systems were superior to their predecessors, nevertheless they still provided only limited modeling capabilities for mission-driven configurations. For example, constraint-based systems are adequate for assembling a bike from an existing catalogue of components and assembling on-the-fly a valid bicycle structure based on the components. However, if dealing with a professional cyclist requiring a completely customized bike suited to his weight, height, to the road surface and climatic conditions, then the constraint-based configurator would prove inadequate. The configurator would be unable to compute the effect of variables such as weight, speed and surface on road friction and drag coefficient and the consequent structure of the bike for optimum performance. In the absence of mission-driven modeling, companies have attempted to develop in-house written models using third party languages. These systems either operate as islands of information or the developer must construct complex integration code between the two systems. Third Generation: Parametric Configurators - Generative Technology In response to the limitations of existing systems, a new generation of technology has emerged that meets the functional demands of today’s sales configuration system users. Parametric configuration has full capacity to support mission-driven configuration, interactive sales configuration and automated output generation. Generative technology uses a single object-oriented language to represent the full spectrum of configuration logic such as bi-directional component constraints and connections, performance calculations, functional definitions, exception rules, geometric modeling and product outputs. Generative languages express a product’s true "what, how and why" description. They use a high level of abstraction to describe a product’s form, fit and function which results in extremely concise and readable models. This streamlined approach makes application development straightforward and keeps maintenance to a minimum. This robust configuration logic can configure products based on inputs about how the product will be used and the environment in which it will be used. Generative systems also excel in their ability to produce automated outputs. Reports, proposals, product visualizations, engineering drawings, bills of material, assembly instructions, routings and other manufacturing information can be automatically generate "on the fly". |
|
Send mail to cjenkin@emory.edu with questions or comments about this web site.
|
