Solution techniques for constraint satisfaction problems: Foundations

The Constraint Satisfaction Problem (CSP) is ubiquitous in artificial intelligence. It has a wide applicability, ranging from machine vision and temporal reasoning to planning and logic programming. This paper attempts a systematic and coherent review of the foundations of the techniques for constraint satisfaction. It discusses in detail the fundamental principles and approaches. This includes an initial definition of the constraint satisfaction problem, a graphical means of problem representation, conventional tree search solution techniques, and pre-processing algorithms which are designed to make subsequent tree search significantly easier.