Complexity theory aims to understand and classify computational problems, especially decision problems, according to their inherent complexity. This book uses new techniques to expand the theory for use with counting problems. The authors present dichotomy classifications for broad classes of counting problems in the realm of P and NP. Classifications are proved for partition functions of spin systems, graph homomorphisms, constraint satisfaction problems, and Holant problems. The book assumes minimal prior knowledge of computational complexity theory, developing proof techniques as needed and gradually increasing the generality and abstraction of the theory. This volume presents the theory on the Boolean domain, and includes a thorough presentation of holographic algorithms, culminating in classifications of computational problems studied in exactly solvable models from statistical mechanics.

Complexity theory aims to understand and classify computational problems, especially decision problems, according to their inherent complexity. This book uses new techniques to expand the theory for use with counting problems. The authors present dichotomy classifications for broad classes of counting problems in the realm of P and NP. Classifications are proved for partition functions of spin systems, graph homomorphisms, constraint satisfaction problems, and Holant problems. The book assumes minimal prior knowledge of computational complexity theory, developing proof techniques as needed and gradually increasing the generality and abstraction of the theory. This volume presents the theory on the Boolean domain, and includes a thorough presentation of holographic algorithms, culminating in classifications of computational problems studied in exactly solvable models from statistical mechanics

Complexity theory aims to understand and classify computational problems, especially decision problems, according to their inherent complexity. This book uses new techniques to expand the theory for use with counting problems. The authors present dichotomy classifications for broad classes of counting problems in the realm of P and NP. Classifications are proved for partition functions of spin systems, graph homomorphisms, constraint satisfaction problems, and Holant problems. The book assumes minimal prior knowledge of computational complexity theory, developing proof techniques as needed and gradually increasing the generality and abstraction of the theory. This volume presents the theory on the Boolean domain, and includes a thorough presentation of holographic algorithms, culminating in classifications of computational problems studied in exactly solvable models from statistical mechanics.

A sweeping classification theory for computational counting problems using new techniques and theories.

Volume 1. Boolean domain

This two-part volume represents the proceedings of the Fifth International Congress of Chinese Mathematicians, held at Tsinghua University, Beijing, in December 2010. The Congress brought together eminent Chinese and overseas mathematicians to discuss the latest developments in pure and applied mathematics. Included are 60 papers based on lectures given at the conference.

We study the computational complexity of counting problems defined over graphs. Complexity dichotomies are proved for various sets of problems, which classify the complexity of each problem in the set as either computable in polynomial time or \#P-hard. These problems are expressible in the frameworks of counting graph homomorphisms, counting constraint satisfaction problems, or Holant problems. However, the proofs are always expressed within the framework of Holant problems, which contains the other two frameworks as special cases. Holographic transformations are naturally expressed using this framework. They represent proofs that two different-looking problems are actually the same. We use them to prove both hardness and tractability. Moreover, the tractable cases are often stated using a holographic transformation. The uniting theme in the proofs of every dichotomy is the technical advances achieved in order to prove the hardness. Specifically, polynomial interpolation appears prominently and is indispensable. We repeatedly strengthen and extend this technique and are rewarded with dichotomies for larger and larger classes of problems. We now have a thorough understanding of its power as well as its ultimate limitations. However, fundamental questions remain since polynomial interpolation is intimately connected with integer solutions of algebraic curves and determinations of Galois groups, subjects that remain active areas of research in pure mathematics. Our motivation for this work is to understand the limits of efficient computation. Without settling the P versus #P question, the best hope is to achieve such complexity classifications.