A COMPARATIVE STUDY OF GENERALIZED ARC-CONSISTENCY ALGORITHMS

Authors

Olufikayo S. Adetunji, University of Nebraska – LincolnFollow

Date of this Version

12-2014

Comments

A THESIS Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Master of Science, Major: Computer Science, Under the Supervision of Professor Berthe Y. Choueiry. Lincoln, Nebraska: December, 2014

Copyright 2014 Olufikayo S. Adetunji

Abstract

In this thesis, we study several algorithms for enforcing Generalized Arc­-Consistency (GAC), which is the most popular consistency property for solving Constraint Satisfaction Problems (CSPs) with backtrack search. The popularity of such algorithms stems from their relative low cost and effectiveness in improving the performance of search. Virtually all commercial and public- domain constraint solvers include some implementation of a generic GAC algorithm. In recent years, several algorithms for enforcing GAC have been proposed in the literature that relies on increasingly complex data structures and mechanisms to improve performance. In this thesis, we study, assess, and compare a basic algorithm for generic constraints (i.e., GAC2001), new algorithms for table constraints (i.e., STR1, STR2, STR3, eSTR1, eSTR2, and STR­Ni), and an algorithm for constraints expressed as multi­valued decision diagram (i.e., mddc). We explain the mechanisms of the above algorithms, and empirically evaluate and compare their performances. We propose a new hybrid algorithm that uses a selection criterion to combine the use of STR1 and STR­Ni.

Adviser: Berthe Y. Choueiry