Abstract
Modelers can effectively use explanations as a suitable method to test, debug, and maintain a model of a product configuration problem. They can use explanations to identify the inconsistencies that modeling errors might cause as such errors can be difficult to isolate if constraint sets are large and the constraints are complex. The explanations for an expert might be presented in terms of his or her own choices, and also in terms of constraints within the configuration model. Truth-maintenace systems (TMS) offer a mechanism that has been used extensively as a basis for generating explanations in product configurators. A TMS is defined by a set of facts, including the atomic facts 'true' and 'false', and a set of jurisdictions, and it treats each fact as a logical proposition. QuickXplain also speeds up the basic iterative testing scheme using a divide-and-conquer approach and is logarithmic in the number of foreground constraints if the explanations are small.
| Original language | English |
|---|---|
| Pages (from-to) | 83-85 |
| Number of pages | 3 |
| Journal | IEEE Intelligent Systems |
| Volume | 22 |
| Issue number | 1 |
| DOIs | |
| Publication status | Published - Jan 2007 |