Problem Solving in Open Worlds

This book presents a computational model of design problem solving in which problems may be under-specified and domain knowledge may be incomplete. The approach combines plausible inference methods for generating, adapting, and evaluating designs, together with a flexible control strategy that accommodates constraints emerging late in the design process.

Novel features of this model include a problem-solving architecture that integrates case-based reasoning and constraint posting to generate plausible designs, domain-independent algorithms and heuristics for adapting proposed designs, and new mechanisms and metrics for evaluating solutions in terms of their integrity and completeness. This model of design is implemented in a computer program called JULIA that interactively designs the presentation and menu of a meal.

Aceasta carte prezinta un model de calcul al rezolvarii problemelor de proiectare in care problemele pot fi subspecificate si cunostintele de domeniu pot fi incomplete. Abordarea combina metode de inferenta plauzibile pentru generarea, adaptarea si evaluarea proiectelor, impreuna cu o strategie de control flexibila care sa acomodeze constrangerile care apar tarziu in procesul de proiectare. Caracteristicile noi ale acestui model includ o arhitectura de rezolvare a problemelor care integreaza rationamentul si constrangerile bazate pe cazuri. postarea pentru a genera modele plauzibile, algoritmi si euristici independente de domeniu pentru adaptarea proiectelor propuse si noi mecanisme si valori pentru evaluarea solutiilor in ceea ce priveste integritatea si completitudinea acestora.