EP 0.82

S. Schulz
Institut für Informatik, Technische Universität München, Germany, and ITC/irst, Trento, Italy
schulz@informatik.tu-muenchen.de

EP will use the current version of the new TSTP output format, documented in [SZS2003]. Our current implementation lists initial clauses as of "unknown" origin. The following rule names are defined for the main proof search:

er
Equality resolution: x!=a v x=x ==> a=a
pm
Paramodulation. Note that E considers all literals as equational, and thus also performs resolution by a combination of top-level paramodulation and (implicit) clause normalization.
ef
Equality factoring (factor equations on one side only, move the remaining disequation into the precondition): x=a v b=c v x=d ==> a!=c v b=c vb=d
split
Clause splitting a la Vampire (non-deductive, but maintains unsatisfiability)
rw
Rewriting, each rw-expression corresponds to exacly one rewrite step with the named clause. This is also used for equational unfolding.
sr
Simplify-reflect: An (equational) version of unit-cutting. As as example, see this positive simplify-reflect step: [a=b], [f(a)!=f(b)] => [].
csr
Contextual simplify-reflect (also known as contextual literal cutting or subsumption resolution): Literal cutting with non-unit clauses if one of the literals is implied in the context of the clause to be simplified.
ar
AC-resolution: Delete literals that are trivial modulo the AC-theory induced by the named clauses
cn
Clause normalize, delete trivial and repeated literals. This step is often implicit, but can sometimes occcur explicitely.

Additionally, the clausification will use additional rule names:

assume_negation
Negate a conjecture for a proof by refutation.
fof_nnf
Convert a formula to negation normal form by moving negation signs inward to the literals, and eleminating equivalences and implications.
shift_quantors
Move quantors (inwards for mini-scoping or outwards for the final conjunctive normal form).
variable_rename
Rename bound variables to make sure each quantor binds a different variable.
skolemize
Eliminate existential quantors via Skolemization.
distribute
Apply distributivity law to move and outwards over or.
split_conjunct
Split of one conjunct from a formula in conjunctive normal form to create a clause.
fof_simplification
Perform standard simplification steps on the formula. This may also replaces back implications and exclusive or with equivalent constructs using implications and equivalences.

The first proof uses all proof search inferences except for "ef" and "csr", although it uses some in fairly trivial ways. The second is the required proof for SYN075-1, and contains examples for "ef" and "csr". The final proof is for SYN075+1, and also contains the clausification steps.

ALL_RULES

# Problem is unsatisfiable, constructing proof object
# TSTP exit status: Unsatisfiable
# Proof object starts here.
cnf(1,axiom,(equal(f(X1,X2), f(X2,X1))),unknown).
cnf(2,axiom,(equal(f(X1,f(X2,X3)), f(f(X1,X2),X3))),unknown).
cnf(3,axiom,(equal(g(X1,X2), g(X2,X1))),unknown).
cnf(4,axiom,(~equal(f(f(X1,X2),f(X3,g(X4,X5))), f(f(g(X4,X5),X3),f(X2,X1)))|~equal(k(X1,X1), k(a,b))),unknown).
cnf(5,axiom,(equal(b, c)|~equal(X1, X2)|~equal(X3, X4)|~equal(c, d)),unknown).
cnf(6,axiom,(equal(a, b)|equal(a, c)),unknown).
cnf(7,axiom,(equal(i(X1), i(X2))),unknown).
cnf(8,axiom,(equal(c, d)|~equal(h(i(a)), h(i(e)))),unknown).
cnf(13,plain-derived,(~equal(k(a,b), k(X1,X1))),inference(ar,[status(thm)],[4,1,3,2,theory(equality)])).
cnf(23,plain-derived,(equal(c, b)|epred1_0|~equal(d, c)|~equal(X3, X4)),inference(split,[split(esplit,[])],[5])).
cnf(24,plain-derived,(epred2_0|~equal(X1, X2)),inference(split,[split(esplit,[])],[5])).
cnf(25,plain-derived,(~epred2_0|~epred1_0),inference(split,[split(esplit,[])],[5])).
cnf(26,plain-derived,(epred2_0),inference(er,[status(thm)],[24,theory(equality)])).
cnf(27,plain-derived,(false|~epred1_0),inference(rw,[status(thm)],[25,26,theory(equality)])).
cnf(28,plain-derived,(~epred1_0),inference(cn,[status(thm)],[27,theory(equality)])).
cnf(29,plain-derived,(equal(c, b)|epred1_0|~equal(d, c)),inference(er,[status(thm)],[23,theory(equality)])).
cnf(30,plain-derived,(equal(c, b)|~equal(d, c)),inference(sr,[status(thm)],[29,28,theory(equality)])).
cnf(31,plain-derived,(equal(d, c)),inference(sr,[status(thm)],[8,7,theory(equality)])).
cnf(32,plain-derived,(equal(c, b)|false),inference(rw,[status(thm)],[30,31,theory(equality)])).
cnf(33,plain-derived,(equal(c, b)),inference(cn,[status(thm)],[32,theory(equality)])).
cnf(34,plain-derived,(equal(b, a)),inference(pm,[status(thm)],[6,33,theory(equality)])).
cnf(36,plain-derived,(~equal(k(b,b), k(X1,X1))),inference(rw,[status(thm)],[13,34,theory(equality)])).
cnf(75,plain-derived,(false),inference(er,[status(thm)],[36,theory(equality)])).
cnf(76,plain-derived,(false),75,['proof']).
# Proof object ends here.

SYN075-1

# Problem is unsatisfiable, constructing proof object
# TSTP exit status: Unsatisfiable
# Proof object starts here.
cnf(1,axiom,(equal(X1, a)|~big_f(X1,X2)),unknown).
cnf(3,axiom,(big_f(X1,X2)|~equal(X1, a)|~equal(X2, b)),unknown).
cnf(4,conjecture,(equal(f(X2), X2)|~big_f(X1,f(X2))|~equal(X1, g(X2))),unknown).
cnf(6,conjecture,(big_f(X1,f(X2))|equal(f(X2), X2)|~equal(X1, g(X2))),unknown).
cnf(9,conjecture,(big_f(h(X1,X2),f(X1))|equal(h(X1,X2), X2)|~equal(f(X1), X1)),unknown).
cnf(10,conjecture,(~equal(f(X1), X1)|~equal(h(X1,X2), X2)|~big_f(h(X1,X2),f(X1))),unknown).
cnf(13,conjecture-derived,(~equal(f(X1), X1)|~equal(h(X1,X2), X2)|~equal(a, h(X1,X2))|~equal(b, f(X1))),inference(pm,[status(thm)],[10,3,theory(equality)])).
cnf(15,conjecture-derived,(equal(f(X2), X2)|~equal(g(X2), X1)),inference(csr,[status(thm)],[6,4])).
cnf(16,conjecture-derived,(equal(f(X1), X1)),inference(er,[status(thm)],[15,theory(equality)])).
cnf(23,conjecture-derived,(false|~equal(h(X1,X2), X2)|~equal(h(X1,X2), a)|~equal(f(X1), b)),inference(rw,[status(thm)],[13,16,theory(equality)])).
cnf(24,conjecture-derived,(false|~equal(h(X1,X2), X2)|~equal(h(X1,X2), a)|~equal(X1, b)),inference(rw,[status(thm)],[23,16,theory(equality)])).
cnf(25,conjecture-derived,(~equal(h(X1,X2), X2)|~equal(h(X1,X2), a)|~equal(X1, b)),inference(cn,[status(thm)],[24,theory(equality)])).
cnf(30,conjecture-derived,(equal(h(X1,X2), X2)|big_f(h(X1,X2),X1)|~equal(f(X1), X1)),inference(rw,[status(thm)],[9,16,theory(equality)])).
cnf(31,conjecture-derived,(equal(h(X1,X2), X2)|big_f(h(X1,X2),X1)|false),inference(rw,[status(thm)],[30,16,theory(equality)])).
cnf(32,conjecture-derived,(equal(h(X1,X2), X2)|big_f(h(X1,X2),X1)),inference(cn,[status(thm)],[31,theory(equality)])).
cnf(38,conjecture-derived,(equal(a, h(X1,X2))|equal(h(X1,X2), X2)),inference(pm,[status(thm)],[1,32,theory(equality)])).
cnf(51,conjecture-derived,(equal(h(X5,X6), a)|~equal(X6, a)),inference(ef,[status(thm)],[38,theory(equality)])).
cnf(54,conjecture-derived,(equal(h(X1,X2), X2)|~equal(a, X2)|~equal(h(X1,X2), a)|~equal(X1, b)),inference(pm,[status(thm)],[25,38,theory(equality)])).
cnf(85,conjecture-derived,(~equal(a, X2)|~equal(h(X1,X2), a)|~equal(X1, b)),inference(csr,[status(thm)],[54,25])).
cnf(86,conjecture-derived,(~equal(a, X2)|~equal(X1, b)),inference(csr,[status(thm)],[85,51])).
cnf(87,conjecture-derived,(epred1_0|~equal(a, X2)),inference(split,[split(esplit,[])],[86])).
cnf(88,conjecture-derived,(epred2_0|~equal(X1, b)),inference(split,[split(esplit,[])],[86])).
cnf(89,conjecture-derived,(~epred2_0|~epred1_0),inference(split,[split(esplit,[])],[86])).
cnf(90,conjecture-derived,(epred1_0),inference(er,[status(thm)],[87,theory(equality)])).
cnf(92,conjecture-derived,(~epred2_0|false),inference(rw,[status(thm)],[89,90,theory(equality)])).
cnf(93,conjecture-derived,(~epred2_0),inference(cn,[status(thm)],[92,theory(equality)])).
cnf(94,conjecture-derived,(~equal(X1, b)),inference(sr,[status(thm)],[88,93,theory(equality)])).
cnf(95,conjecture-derived,(false),inference(er,[status(thm)],[94,theory(equality)])).
cnf(97,conjecture-derived,(false),95,['proof']).
# Proof object ends here.

SYN075+1

# Problem is unsatisfiable, constructing proof object
# TSTP exit status: Unsatisfiable
# Proof object starts here.
fof(1, axiom,?[X1]:?[X2]:![X3]:![X4]:(big_f(X3,X4)<=>(equal(X3, X1)&equal(X4, X2))),unknown).
fof(2, conjecture,?[X2]:![X4]:(?[X1]:![X3]:(big_f(X3,X4)<=>equal(X3, X1))<=>equal(X4, X2)),unknown).
fof(3, assumption-derived,~(?[X2]:![X4]:(?[X1]:![X3]:(big_f(X3,X4)<=>equal(X3, X1))<=>equal(X4, X2))),inference(assume_negation,[status(cth)],[2])).
fof(4, plain-derived,?[X1]:?[X2]:![X3]:![X4]:((~(big_f(X3,X4))|(equal(X3, X1)&equal(X4, X2)))&((~(equal(X3, X1))|~(equal(X4, X2)))|big_f(X3,X4))),inference(fof_nnf,[status(thm)],[1])).
fof(5, plain-derived,?[X1]:?[X2]:(![X3]:![X4]:(~(big_f(X3,X4))|(equal(X3, X1)&equal(X4, X2)))&![X3]:![X4]:((~(equal(X3, X1))|~(equal(X4, X2)))|big_f(X3,X4))),inference(shift_quantors,[status(thm)],[4])).
fof(6, plain-derived,?[X5]:?[X6]:(![X7]:![X8]:(~(big_f(X7,X8))|(equal(X7, X5)&equal(X8, X6)))&![X9]:![X10]:((~(equal(X9, X5))|~(equal(X10, X6)))|big_f(X9,X10))),inference(variable_rename,[status(thm)],[5])).
fof(7, plain-derived,(![X7]:![X8]:(~(big_f(X7,X8))|(equal(X7, esk1_0)&equal(X8, esk2_0)))&![X9]:![X10]:((~(equal(X9, esk1_0))|~(equal(X10, esk2_0)))|big_f(X9,X10))),inference(skolemize,[status(sab)],[6])).
fof(8, plain-derived,![X7]:![X8]:![X9]:![X10]:(((equal(X7, esk1_0)|~(big_f(X7,X8)))&(equal(X8, esk2_0)|~(big_f(X7,X8))))&((~(equal(X9, esk1_0))|~(equal(X10, esk2_0)))|big_f(X9,X10))),inference(distribute,[status(thm)],[7])).
cnf(9,plain-derived,(big_f(X9,X10)|~equal(X10, esk2_0)|~equal(X9, esk1_0)),inference(split_conjunct,[status(thm)],[8])).
cnf(10,plain-derived,(equal(X8, esk2_0)|~big_f(X7,X8)),inference(split_conjunct,[status(thm)],[8])).
cnf(11,plain-derived,(equal(X7, esk1_0)|~big_f(X7,X8)),inference(split_conjunct,[status(thm)],[8])).
fof(12, assumption-derived,![X2]:?[X4]:((![X1]:?[X3]:((~(big_f(X3,X4))|~(equal(X3, X1)))&(big_f(X3,X4)|equal(X3, X1)))|~(equal(X4, X2)))&(?[X1]:![X3]:((big_f(X3,X4)&equal(X3, X1))|(~(big_f(X3,X4))&~(equal(X3, X1))))|equal(X4, X2))),inference(fof_nnf,[status(thm)],[3])).
fof(13, assumption-derived,![X5]:?[X6]:((![X7]:?[X8]:((~(big_f(X8,X6))|~(equal(X8, X7)))&(big_f(X8,X6)|equal(X8, X7)))|~(equal(X6, X5)))&(?[X9]:![X10]:((big_f(X10,X6)&equal(X10, X9))|(~(big_f(X10,X6))&~(equal(X10, X9))))|equal(X6, X5))),inference(variable_rename,[status(thm)],[12])).
fof(14, assumption-derived,![X5]:((![X7]:((~(big_f(esk4_2(X5,X7),esk3_1(X5)))|~(equal(esk4_2(X5,X7), X7)))&(big_f(esk4_2(X5,X7),esk3_1(X5))|equal(esk4_2(X5,X7), X7)))|~(equal(esk3_1(X5), X5)))&(![X10]:((big_f(X10,esk3_1(X5))&equal(X10, esk5_1(X5)))|(~(big_f(X10,esk3_1(X5)))&~(equal(X10, esk5_1(X5)))))|equal(esk3_1(X5), X5))),inference(skolemize,[status(sab)],[13])).
fof(15, assumption-derived,![X5]:![X7]:![X10]:((((~(big_f(esk4_2(X5,X7),esk3_1(X5)))|~(equal(esk4_2(X5,X7), X7)))|~(equal(esk3_1(X5), X5)))&((big_f(esk4_2(X5,X7),esk3_1(X5))|equal(esk4_2(X5,X7), X7))|~(equal(esk3_1(X5), X5))))&((((~(big_f(X10,esk3_1(X5)))|big_f(X10,esk3_1(X5)))|equal(esk3_1(X5), X5))&((~(equal(X10, esk5_1(X5)))|big_f(X10,esk3_1(X5)))|equal(esk3_1(X5), X5)))&(((~(big_f(X10,esk3_1(X5)))|equal(X10, esk5_1(X5)))|equal(esk3_1(X5), X5))&((~(equal(X10, esk5_1(X5)))|equal(X10, esk5_1(X5)))|equal(esk3_1(X5), X5))))),inference(distribute,[status(thm)],[14])).
cnf(18,conjecture-derived,(equal(esk3_1(X5), X5)|big_f(X10,esk3_1(X5))|~equal(X10, esk5_1(X5))),inference(split_conjunct,[status(thm)],[15])).
cnf(20,conjecture-derived,(equal(esk4_2(X5,X7), X7)|big_f(esk4_2(X5,X7),esk3_1(X5))|~equal(esk3_1(X5), X5)),inference(split_conjunct,[status(thm)],[15])).
cnf(21,conjecture-derived,(~equal(esk3_1(X5), X5)|~equal(esk4_2(X5,X7), X7)|~big_f(esk4_2(X5,X7),esk3_1(X5))),inference(split_conjunct,[status(thm)],[15])).
cnf(24,conjecture-derived,(~equal(esk3_1(X1), X1)|~equal(esk4_2(X1,X2), X2)|~equal(esk1_0, esk4_2(X1,X2))|~equal(esk2_0, esk3_1(X1))),inference(pm,[status(thm)],[21,9,theory(equality)])).
cnf(26,conjecture-derived,(equal(esk3_1(X1), X1)|big_f(esk5_1(X1),esk3_1(X1))),inference(er,[status(thm)],[18,theory(equality)])).
cnf(27,conjecture-derived,(equal(esk1_0, esk4_2(X1,X2))|equal(esk4_2(X1,X2), X2)|~equal(esk3_1(X1), X1)),inference(pm,[status(thm)],[11,20,theory(equality)])).
cnf(28,conjecture-derived,(equal(esk2_0, esk3_1(X1))|equal(esk4_2(X1,X2), X2)|~equal(esk3_1(X1), X1)),inference(pm,[status(thm)],[10,20,theory(equality)])).
cnf(31,conjecture-derived,(equal(esk1_0, esk5_1(X1))|equal(esk3_1(X1), X1)),inference(pm,[status(thm)],[11,26,theory(equality)])).
cnf(32,conjecture-derived,(equal(esk2_0, esk3_1(X1))|equal(esk3_1(X1), X1)),inference(pm,[status(thm)],[10,26,theory(equality)])).
cnf(35,conjecture-derived,(big_f(esk1_0,esk3_1(X1))|equal(esk3_1(X1), X1)),inference(pm,[status(thm)],[26,31,theory(equality)])).
cnf(36,conjecture-derived,(equal(esk3_1(X2), X2)|~equal(esk2_0, X2)),inference(ef,[status(thm)],[32,theory(equality)])).
cnf(50,conjecture-derived,(~big_f(esk4_2(X1,X2),X1)|~equal(esk3_1(X1), X1)|~equal(esk4_2(X1,X2), X2)|~equal(esk2_0, X1)),inference(pm,[status(thm)],[21,36,theory(equality)])).
cnf(64,conjecture-derived,(big_f(esk1_0,esk2_0)|equal(esk3_1(X1), X1)),inference(pm,[status(thm)],[35,32,theory(equality)])).
cnf(79,conjecture-derived,(~big_f(esk4_2(X1,X2),X1)|~equal(esk4_2(X1,X2), X2)|~equal(esk2_0, X1)),inference(csr,[status(thm)],[50,36])).
cnf(80,conjecture-derived,(~big_f(esk4_2(X1,X2),X1)|~equal(esk4_2(X1,X2), X2)),inference(csr,[status(thm)],[79,10])).
cnf(100,conjecture-derived,(equal(esk3_1(X1), esk2_0)|equal(esk4_2(X1,X2), X2)),inference(csr,[status(thm)],[28,32])).
cnf(104,conjecture-derived,(equal(esk3_1(X1), esk2_0)|~big_f(X2,X1)|~equal(esk4_2(X1,X2), X2)),inference(pm,[status(thm)],[80,100,theory(equality)])).
cnf(108,conjecture-derived,(equal(esk3_1(X1), esk2_0)|~big_f(X2,X1)),inference(csr,[status(thm)],[104,100])).
cnf(113,conjecture-derived,(equal(esk3_1(esk2_0), esk2_0)|equal(esk3_1(X1), X1)),inference(pm,[status(thm)],[108,64,theory(equality)])).
cnf(136,conjecture-derived,(equal(esk3_1(esk2_0), esk2_0)),inference(ef,[status(thm)],[113,theory(equality)])).
cnf(174,conjecture-derived,(equal(esk4_2(esk2_0,X1), esk1_0)|equal(esk4_2(esk2_0,X1), X1)),inference(pm,[status(thm)],[27,136,theory(equality)])).
cnf(250,conjecture-derived,(equal(esk4_2(esk2_0,X1), X1)|~equal(esk1_0, X1)),inference(ef,[status(thm)],[174,theory(equality)])).
cnf(291,conjecture-derived,(~equal(esk3_1(esk2_0), esk2_0)|~equal(esk4_2(esk2_0,X1), esk1_0)|~equal(esk1_0, X1)),inference(pm,[status(thm)],[24,250,theory(equality)])).
cnf(304,conjecture-derived,(false|~equal(esk4_2(esk2_0,X1), esk1_0)|~equal(esk1_0, X1)),inference(rw,[status(thm)],[291,136,theory(equality)])).
cnf(305,conjecture-derived,(~equal(esk4_2(esk2_0,X1), esk1_0)|~equal(esk1_0, X1)),inference(cn,[status(thm)],[304,theory(equality)])).
cnf(309,conjecture-derived,(~equal(X1, esk1_0)),inference(pm,[status(thm)],[305,250,theory(equality)])).
cnf(311,conjecture-derived,(false),inference(er,[status(thm)],[309,theory(equality)])).
cnf(314,conjecture-derived,(false),311,['proof']).
# Proof object ends here.

References

SZS2003
Suctcliffe, G.; Zimmer, J.; Schulz S. (2003), Communication Formalisms for Automated Theorem Proving Tools, Proceedings of the IJCAI-18 Workshop on Agents and Automated Reasoning, (Acapulco, Mexico).