Information about Automated Theorem Proving

Automated theorem proving (ATP) or automated deduction, currently the most well-developed subfield of automated reasoning (AR), is the proving of mathematical theorems by a computer program.

Decidability of the problem

Depending on the underlying logic, the problem of deciding the validity of a theorem varies from trivial to impossible. For the frequent case of propositional logic, the problem is decidable but NP-complete, and hence only exponential-time algorithms are believed to exist for general proof tasks. For a first order predicate calculus, that is having no proper axioms, Gödel's Completeness Theorem states that the theorems are exactly the logically valid well-formed formulas, so identifying theorems is recursively enumerable, i.e., given unbounded resources, any valid theorem can eventually be proven. Invalid statements, i.e. formulas that are not entailed by a given theory, cannot always be recognized. In addition, a consistent formal theory that contains the first-order theory of the natural numbers (having certain proper axioms then), by Gödel's incompleteness theorems, contains a true statement which cannot be proven, in which case a theorem prover trying to prove such a statement ends up in nontermination.

In these cases, a first-order theorem prover may fail to terminate while searching for a proof. Despite these theoretical limits, practical theorem provers can solve many hard problems in these logics.

Related problems

A simpler, but related problem is proof verification, where an existing proof for a theorem is certified valid. For this, it is generally required that each individual proof step can be verified by a primitive recursive function or program, and hence the problem is always decidable.

Interactive theorem provers require a human user to give hints to the system. Depending on the degree of automation, the prover can essentially be reduced to a proof checker, with the user providing the proof in a formal way, or significant proof tasks can be performed automatically. Interactive provers are used for a variety of tasks, but even fully automatic systems have by now proven a number of interesting and hard theorems, including some that have eluded human mathematicians for a long time. However, these successes are sporadic, and work on hard problems usually requires a proficient user.

Another distinction is sometimes drawn between theorem proving and other techniques, where a process is considered to be theorem proving if it consists of a traditional proof, starting with axioms and producing new inference steps using rules of inference. Other techniques would include model checking, which is equivalent to brute-force enumeration of many possible states (although the actual implementation of model checkers requires much cleverness, and does not simply reduce to brute force). There are hybrid theorem proving systems which use model checking as an inference rule. There are also programs which were written to prove a particular theorem, with a (usually informal) proof that if the program finishes with a certain result, then the theorem is true. A good example of this was the machine-aided proof of the four color theorem, which was very controversial as the first claimed mathematical proof which was essentially impossible to verify by humans due to the enormous size of the program's calculation (such proofs are called non-surveyable proofs). Another example would be the proof that the game Connect Four is a win for the first player.

Industrial uses

Commercial use of automated theorem proving is mostly concentrated in integrated circuit design and verification. Since the Pentium FDIV bug, the complicated floating point units of modern microprocessors have been designed with extra scrutiny. In the latest processors from AMD, Intel, and others, automated theorem proving has been used to verify that division and other operations are correct.

First-order theorem proving

So-called first-order theorem proving may be restricted to a propositional calculus with terms (constants, function names, and free variables) added, making it impossible to express mathematical induction. It should then not be confused with a first-order theory of metamathematics, as the quantifiers have been stripped out, though universal quantifiers may be emulated by rewriting into free variables.

First-order theorem proving is one of the most mature subfields of automated theorem proving. The logic is expressive enough to allow the specification of arbitrary problems, often in a reasonably natural and intuitive way. On the other hand, it is still semi-decidable, and a number of sound and complete calculi have been developed, enabling fully automated systems. More expressive logics, such as higher order and modal logics, allow the convenient expression of a wider range of problems than first order logic, but theorem proving for these logics is less well developed. The quality of implemented system has benefited by the existence of a large library of standard benchmark examples (the TPTP), as well as by the CADE ATP System Competition (CASC), a yearly competition of first-order systems for many important classes of first-order problems.

Some important systems (all have won at least one CASC competition division) are listed below.

• E is a high-performance prover for full first-order logic, but built on a purely equational calculus, developed primarily in the automated reasoning group of Technical University of Munich.
• Otter, developed at the Argonne National Laboratory, is the first widely used high-performance theorem prover. It is based on first-order resolution and paramodulation. Otter has since been replaced by Prover9, which is paired with Mace4.
• SETHEO is a high-performance system based on the goal-directed model elimination calculus. It is developed in the automated reasoning group of Technical University of Munich. E and SETHEO have been combined (with other systems) in the composite theorem prover E-SETHEO.
• Vampire is developed and implemented at Manchester University by Andrei Voronkov, formerly together with Alexandre Riazanov. It has won the "world cup for theorem provers" (the CADE ATP System Competition) in the most prestigious CNF (MIX) division for seven years (1999, 2001 - 2006).
• Waldmeister is a specialized system for unit-equational first-order logic. It has won the CASC UEQ division for the last ten years (1997-2006).

Popular techniques

• First-order resolution with unification
• Lean theorem proving
• Model elimination
• Method of analytic tableaux
• Superposition and term rewriting
• Model checking
• Mathematical induction
• Binary decision diagrams
• DPLL
• Higher-order unification

Available implementations

• ACL2
• Acumen RuleManager (commercial product)
• Automath
• Alligator
• Carine
• Coq
• CVC
• E
• Isabelle
• Gandalf
• HOL
• HOL Light
• KeY
• KIV
• LCF
• LoTREC
• MetaPRL
• Matita
• Mizar
• NuPRL
• Otter
• Paradox
• PhoX
• ProofPower
• Prover Plug-In (commercial proof engine product)
• ProverBox
• Prover9/Mace4
• PVS
• ResearchCyc
• Simplify
• SNARK
• SPARK programming language
• SPASS
• Spear modular arithmetic theorem prover
• Tau
• TPS
• Twelf
• Vampire/Vampyre
• Waldmeister

You can find information on some of these theorem provers and others at [1] or the QPQ website. The TPTP library of test problems, suitable for testing first-order theorem provers, is available at [2] and solutions from many of these provers for TPTP problems are in the TSTP solution library, available at [3]

Important people

• Leo Bachmair, co-developer of the superposition calculus.
• Woody Bledsoe, artificial intelligence pioneer.
• Robert S. Boyer, co-author of the Boyer-Moore theorem prover, co-recipient of the Herbrand Award 1999.
• Alan Bundy, University of Edinburgh, meta-level reasoning for guiding inductive proof, proof planning and recipient of 2007 IJCAI Award for Research Excellence and 2003 Donald E. Walker Distinguished Service Award.
• William McCune Argonne National Laboratory, author of Otter, the first high-performance theorem prover. Many important papers, recipient of the Herbrand Award 2000.
• Robert Constable, Cornell University. Important contributions to type theory, NuPRL.
• Martin Davis, author of the "Handbook of Artificial Reasoning", co-inventor of the DPLL algorithm, recipient of the Herbrand Award 2005.
• Branden Fitelson University of California at Berkeley. Work in shortest axiomatic bases for logic systems.
• Harald Ganzinger, co-developer of the superposition calculus, head of the MPI Saarbrücken, recipient of the Herbrand Award 2004 (posthumous).
• Michael Genesereth, Stanford University professor of Computer Science.
• Keith Goolsbey chief developer of the Cyc inference engine.
• Michael J. C. Gordon led the development of the HOL theorem prover.
• Robert Kowalski developed the connection graph theorem-prover and SLD resolution, the inference engine that executes logic programs.
• Donald W. Loveland Duke University. Author, co-developer of the DPLL-procedure, developer of model elimination, recipient of the Herbrand Award 2001.
• Sergei Maslov
• Norman Megill, maintainer of metamath.org, an online database of automatically verified proofs.
• J Strother Moore, co-author of the Boyer-Moore theorem prover, co-recipient of the Herbrand Award 1999.
• Robert Nieuwenhuis University of Barcelona. Co-developer of the superposition calculus.
• Tobias Nipkow Technical University of Munich, contributions to (higher-order) rewriting, co-developer of the Isabelle, proof assistant
• Ross Overbeek Argonne National Laboratory. Founder of The Fellowship for Interpretation of Genomes
• Lawrence C. Paulson University of Cambridge, work on higher-order logic system, co-developer of the Isabelle proof assistant
• David A. Plaisted University of North Carolina at Chapel Hill. Complexity results, contributions to rewriting and completion, instance-based theorem proving.
• John Rushby Program Director - SRI International
• J. Alan Robinson Syracuse University. Developed original resolution and unification based first order theorem proving, co-editor of the "Handbook of Automated Reasoning", recipient of the Herbrand Award 1996
• Natarajan Shankar SRI International, work on decision procedures, little engines of proof, co-developer of PVS.
• Mark Stickel SRI. Recipient of the Herbrand Award 2002.
• Geoff Sutcliffe University of Miami. Maintainer of the TPTP collection, an organizer of the CADE annual contest.
• Robert Veroff University of New Mexico. Many important papers.
• Andrei Voronkov Developer of Vampire and Co-Editor of the "Handbook of Automated Reasoning"
• Larry Wos Argonne National Laboratory. (Otter) Many important papers.

References

• Chin-Liang Chang; Richard Char-Tung Lee (1973). Symbolic Logic and Mechanical Theorem Proving. Academic Press. 
• Loveland, Donald W. (1978). Automated Theorem Proving: A Logical Basis. Fundamental Studies in Computer Science Volume 6. North-Holland Publishing. 
• Gallier, Jean H. (1986). Logic for Computer Science: Foundations of Automatic Theorem Proving. Harper & Row Publishers. 
• Duffy, David A. (1991). Principles of Automated Theorem Proving. John Wiley & Sons. 
• Wos, Larry; Overbeek, Ross; Lusk, Ewing; Boyle, Jim (1992). Automated Reasoning: Introduction and Applications, 2nd edition, McGraw-Hill.McGraw-Hill"> 
• (2001) in Alan Robinson and Andrei Voronkov (eds.): Handbook of Automated Reasoning Volume I & II. Elsevier and MIT Press.Elsevier%20and%20MIT%20Press"> 
• Fitting, Melvin (1996). First Order and Automated Theorem Proving, 2nd edition, Springer. 

See also

• Symbolic computation
• Computer-aided proof
• Automated reasoning
• Formal verification
• Logic programming
• Proof checking
• Model checking
• Proof complexity
• Computer algebra system