Added iterative deepening

Printing of Cudd information


git-svn-id: https://yap.svn.sf.net/svnroot/yap/trunk@2271 b08c6af1-5177-4d33-ba66-4b1c6b8b522a

cplint/cplint_yap.c

@@ -134,11 +134,11 @@ static int compute_prob(void)
 	bVar2mVar=array_alloc(int,0);
 	create_dot=YAP_IntOfTerm(arg4);
 	createVars(variables,arg1,mgr,bVar2mVar,create_dot,inames);
-        Cudd_PrintInfo(mgr,stderr);
+        //Cudd_PrintInfo(mgr,stderr);
 
 	/* automatic variable reordering, default method CUDD_REORDER_SIFT used */
-	printf("status %d\n",Cudd_ReorderingStatus(mgr,&order));
+	//printf("status %d\n",Cudd_ReorderingStatus(mgr,&order));
-	printf("order %d\n",order);
+	//printf("order %d\n",order);
 
 	Cudd_AutodynEnable(mgr,CUDD_REORDER_SAME); 
 /*	 Cudd_AutodynEnable(mgr, CUDD_REORDER_RANDOM_PIVOT);

cplint/lpadsld.pl

@@ -21,7 +21,8 @@ setting(ground_body,false).
 if true, both the head and the body of each clause will be grounded, otherwise
 only the head is grounded. In the case in which the body contains variables 
 not appearing in the head, the body represents an existential event */
-
+setting(min_error,0.01).
+setting(depth_bound,4).
 /* end of list of parameters */
 
 /* s(GoalsLIst,Prob) compute the probability of a list of goals 
@@ -98,6 +99,51 @@ solve(GoalsList,Prob,CPUTime1,CPUTime2,WallTime1,WallTime2):-
 	format(user_error,"~nMemory after inference~n",[]),
 	print_mem.
 
+si(GoalsList,ProbL,ProbU,CPUTime):-
+        statistics(cputime,[_,_]),
+	setting(depth_bound,D),
+        solvei(GoalsList,D,ProbL,ProbU),
+	statistics(cputime,[_,CT]),
+	CPUTime is CT/1000.
+
+
+
+solvei(GoalsList,D,ProbL0,ProbU0):-
+	(setof(Deriv,find_deriv(GoalsList,D,Deriv),LDup)->
+		rem_dup_lists(LDup,[],L),
+	%	print_mem,
+		separate_ulb(L,[],LL,[],LU),
+		compute_prob_deriv(LL,ProbL),
+		compute_prob_deriv(LU,ProbU),
+		Err is ProbU-ProbL,
+		setting(min_error,ME),
+		(Err<ME->
+			ProbU0=ProbU,
+			ProbL0=ProbL
+		;
+			setting(depth_bound,DB),
+			D1 is D+DB,
+			solvei(GoalsList,D1,ProbL0,ProbU0)
+		)
+	;
+	%	print_mem,
+		ProbL0=0.0,
+		ProbU0=0.0
+	).
+
+compute_prob_deriv(LL,ProbL):-
+	build_formula(LL,FormulaL,[],VarL,0,ConjL),
+	length(LL,NDL),
+	length(VarL,NVL),
+	%format(user_error,"Disjunctions :~d~nConjunctions: ~d~nVariables ~d~n",[NDL,ConjL,NVL]),
+	var2numbers(VarL,0,NewVarL),
+	(setting(save_dot,true)->
+	%	format("Variables: ~p~n",[VarL]),
+		compute_prob(NewVarL,FormulaL,ProbL,1)
+	;
+		compute_prob(NewVarL,FormulaL,ProbL,0)
+	).
+
 print_mem:-
 	statistics(global_stack,[GS,GSF]),
 	statistics(local_stack,[LS,LSF]),
@@ -111,6 +157,11 @@ print_mem:-
 find_deriv(GoalsList,Deriv):-
 	solve(GoalsList,[],DerivDup),
 	remove_duplicates(DerivDup,Deriv). 
+
+find_deriv(GoalsList,DB,Deriv):-
+        solve(GoalsList,DB,[],DerivDup),
+	remove_duplicates(DerivDup,Deriv).
+
 /* duplicate can appear in the C set because two different unistantiated clauses may become the 
 same clause when instantiated */
 
@@ -146,6 +197,55 @@ solve_cond(Goals,Evidence,Prob):-
 	format(user_error,"~nMemory after inference~n",[]),
 	print_mem.
 
+sci(Goals,Evidence,ProbL,ProbU,CPUTime):-
+	statistics(cputime,[_,_]),
+	setting(depth_bound,D),
+        solve_condi(Goals,Evidence,D,ProbL,ProbU),
+	statistics(cputime,[_,CT]),
+	CPUTime is CT/1000.
+
+solve_condi(Goals,Evidence,D,ProbL0,ProbU0):-
+	(call_residue(setof(DerivE,find_deriv(Evidence,D,DerivE),LDupE),_R0)->
+		rem_dup_lists(LDupE,[],LE),
+		append(Evidence,Goals,EG),
+		(call_residue(setof(DerivGE,find_deriv(EG,D,DerivGE),LDupGE),_R1)->
+			rem_dup_lists(LDupGE,[],LGE),
+			separate_ulb(LGE,[],LLGE,[],LUGE),
+			compute_prob_deriv(LLGE,ProbLGE),
+			compute_prob_deriv(LUGE,ProbUGE),
+			separate_ulb(LE,[],LLE,[],LUE),
+			compute_prob_deriv(LLE,ProbLE),
+			compute_prob_deriv(LUE,ProbUE),
+			ProbL is ProbLGE/ProbUE,
+			(ProbLE=0.0->
+				ProbU1=1.0
+			;
+				ProbU1 is ProbUGE/ProbLE
+			),
+			(ProbU1>1.0->
+				ProbU=1.0
+			;
+				ProbU=ProbU1
+			),
+			Err is ProbU-ProbL,
+			setting(min_error,ME),
+			(Err<ME->
+				ProbU0=ProbU,
+				ProbL0=ProbL
+			;
+				setting(depth_bound,DB),
+				D1 is D+DB,
+				solve_condi(Goals,Evidence,D1,ProbL0,ProbU0)
+			)
+		;
+			ProbL0=0.0,
+			ProbU0=0.0
+		)	
+	;
+		ProbL0=undefined,
+		ProbU0=undefined
+	).
+
 /* sc(Goals,Evidence,Prob,Time1,Time2) compute the conditional probability of the list of goals
 Goals given the list of goals Evidence 
 Goals and Evidence can have variables, sc returns in backtracking all the solutions with their
@@ -216,6 +316,11 @@ find_deriv_GE(LD,GoalsList,Deriv):-
 	solve(GoalsList,D,DerivDup),
 	remove_duplicates(DerivDup,Deriv).
 
+find_deriv_GE(LD,GoalsList,DB,Deriv):-
+        member(D,LD),
+	solve(GoalsList,DB,D,DerivDup),
+	remove_duplicates(DerivDup,Deriv).
+
 /* solve(GoalsList,CIn,COut) takes a list of goals and an input C set
 and returns an output C set
 The C set is a list of triple (N,R,S) where
@@ -295,6 +400,45 @@ solve([H|T],CIn,COut):-
 	find_rule(H,(R,S,N),B,CIn),
 	solve_pres(R,S,N,B,T,CIn,COut).
 
+
+solve([],_DB,C,C):-!.
+
+solve(_G,0,C,[(_,pruned,_)|C]):-!.
+
+solve([\+ H |T],DB,CIn,COut):-!,
+        list2and(HL,H),
+	(setof(D,find_deriv(HL,DB,D),LDup)->
+		rem_dup_lists(LDup,[],L),
+		separate_ulb(L,[],LB,[],UB),
+		(\+ LB=UB->
+			
+			choose_clauses(CIn,LB,C0),
+			C1=[(_,pruned,_)|C0]
+		;
+			choose_clauses(CIn,L,C1)
+		),
+		solve(T,DB,C1,COut)
+	;
+		solve(T,DB,CIn,COut)
+	).
+solve([H|T],DB,CIn,COut):-
+	builtin(H),!,
+	call(H),
+	solve(T,DB,CIn,COut).
+
+solve([H|T],DB,CIn,COut):-
+	def_rule(H,B),
+	append(B,T,NG),
+	DB1 is DB-1,
+	solve(NG,DB1,CIn,COut).
+
+solve([H|T],DB,CIn,COut):-
+	find_rule(H,(R,S,N),B,CIn),
+	DB1 is DB-1,
+	solve_pres(R,S,N,B,T,DB1,CIn,COut).
+
+
+
 solve_pres(R,S,N,B,T,CIn,COut):-
 	member_eq((N,R,S),CIn),!,
 	append(B,T,NG),
@@ -305,6 +449,16 @@ solve_pres(R,S,N,B,T,CIn,COut):-
 	append(B,T,NG),
 	solve(NG,C1,COut).
 
+solve_pres(R,S,N,B,T,DB,CIn,COut):-
+        member_eq((N,R,S),CIn),!,
+	append(B,T,NG),
+	solve(NG,DB,CIn,COut).
+
+solve_pres(R,S,N,B,T,DB,CIn,COut):-
+	append(CIn,[(N,R,S)],C1),
+	append(B,T,NG),
+	solve(NG,DB,C1,COut).
+
 build_initial_graph(N,G):-
 	listN(0,N,Vert),
 	add_vertices([],Vert,G).
@@ -372,6 +526,7 @@ find_rule(H,(R,S,Number),Body,C):-
 
 not_already_present_with_a_different_head(_N,_R,_S,[]).
 
+
 not_already_present_with_a_different_head(N,R,S,[(N1,R,S1)|T]):-
 	not_different(N,N1,S,S1),!,
 	not_already_present_with_a_different_head(N,R,S,T).
@@ -380,6 +535,7 @@ not_already_present_with_a_different_head(N,R,S,[(_N1,R1,_S1)|T]):-
 	R\==R1,
 	not_already_present_with_a_different_head(N,R,S,T).
 
+
 not_different(_N,_N1,S,S1):-
 	S\=S1,!.	
 
@@ -415,6 +571,24 @@ choose_clauses(CIn,[D|T],COut):-
 	impose_dif_cons(R,S,CIn),
 	choose_clauses([(N1,R,S)|CIn],T,COut).
 
+choose_clauses_DB(C,[],C).
+
+choose_clauses_DB(CIn,[D|T],COut):-
+        member((N,R,S),D),
+	ground((N,R,S)),
+	already_present_with_a_different_head(N,R,S,CIn),!,
+	choose_a_head(N,R,S,CIn,C1),
+	choose_clauses_DB(C1,T,COut).
+
+choose_clauses_DB(CIn,[D|T],COut):-
+        member((N,R,S),D),
+	ground((N,R,S)),!,
+	new_head(N,R,S,N1),
+	\+ already_present(N1,R,S,CIn),
+	impose_dif_cons(R,S,CIn),
+	choose_clauses_DB([(N1,R,S)|CIn],T,COut).
+
+
 impose_dif_cons(_R,_S,[]):-!.
 
 impose_dif_cons(R,S,[(_NH,R,SH)|T]):-!,
@@ -511,6 +685,29 @@ member_subset(E,[_H|T]):-
 	member_subset(E,T).
 
 
+separate_ulb([],L,L,U,U):-!.
+/*
+separate_ulb([H|T],L0,L1,U0,[H|U1]):-
+        member(pruned,H),!,
+	separate_ulb(T,L0,L1,U0,U1).
+*/		
+separate_ulb([H|T],L0,[H|L1],U0,[H|U1]):-
+	ground(H),!,
+	separate_ulb(T,L0,L1,U0,U1).
+
+separate_ulb([H|T],L0,L1,U0,[H1|U1]):-
+        get_ground(H,H1),
+	separate_ulb(T,L0,L1,U0,U1).
+
+get_ground([],[]):-!.
+
+get_ground([H|T],[H|T1]):-
+	ground(H),!,
+	get_ground(T,T1).
+
+get_ground([H|T],T1):-
+	get_ground(T,T1).
+
 
 /* predicates for building the formula to be converted into a BDD */
 
@@ -536,8 +733,12 @@ build_formula([D|TD],[F|TF],VarIn,VarOut):-
 	build_term(D,F,VarIn,Var1),
 	build_formula(TD,TF,Var1,VarOut).
 
+
 build_term([],[],Var,Var).
 
+build_term([(_,pruned,_)|TC],TF,VarIn,VarOut):-!,
+	build_term(TC,TF,VarIn,VarOut).
+
 build_term([(N,R,S)|TC],[[NVar,N]|TF],VarIn,VarOut):-
 	(nth0_eq(0,NVar,VarIn,(R,S))->
 		Var1=VarIn

cplint/testlpadsld_gbtrue.pl

@@ -21,6 +21,18 @@ close_to(V,T):-
 	TLow<V,
 	V<THigh.
 
+ti:-
+	format("~nTesting iterative deepening lpadsld.yap~n~n",[]),
+	files(F),
+	statistics(runtime,[_,_]),
+	set(ground_body,true),
+	format("~nGround body~n~n",[]),
+	test_filesi(F,ground_body(true)),
+	statistics(runtime,[_,T]),
+	T1 is T /1000,
+	format("Test successful, time ~f secs.~n",[T1]).
+ti:-
+	format("Test unsuccessful.~n",[]).
 
 t:-
 	format("~nTesting lpadsld.yap~n~n",[]),
@@ -46,6 +58,16 @@ test_files([H|T],GB):-
 	findall(A,test(A,H,GB),L),
 	test_all(H,L),
 	test_files(T,GB).
+test_filesi([],_GB).
+
+test_filesi([H|T],GB):-
+        library_directory(LD),
+	atom_concat(LD,'/cplint/examples/',ExDir),
+	atom_concat(ExDir,H,NH),
+	p(NH),!,
+	findall(A,test(A,H,GB),L),
+	test_alli(H,L),
+	test_filesi(T,GB).
 
 test_all(_F,[]).
 
@@ -56,6 +78,16 @@ test_all(F,[H|T]):-
 	call(H),!,
 	test_all(F,T).
 
+test_alli(_F,[]).
+
+test_alli(F,[H|T]):-
+        copy_term(H,NH),
+	NH=(s(Q,P),close_to('P',_Prob)),
+	format("~a ~p.~n",[F,NH]),
+	si(Q,PL,PU,T),!,
+	format("Lower bound ~f, Upper bound ~f~n",[PL,PU]),
+	test_all(F,T).
+
 
 files([paper_ref_not,paper_ref,female,exapprox,exrange,threesideddice,
 mendel,student,school_simple,school,coin2,ex,trigger,throws,light]).

cplint/testlpadsldit.pl

@@ -0,0 +1,234 @@
+/*
+	LPAD and CP-Logic reasoning suite
+	Copyright (c) 2007, Fabrizio Riguzzi
+
+Test file for lpadsld.pl, case where the body is grounded
+
+Use
+:-t.
+to execute the test
+
+*/
+:-source.
+:-use_module(library(lpadsld)).
+
+
+epsilon(0.000001).
+
+close_to(V,T):-
+	epsilon(E),
+	TLow is T-E,
+	THigh is T+E,
+	TLow<V,
+	V<THigh.
+
+t:-
+	format("~nTesting iterative deepening lpadsld.yap~n~n",[]),
+	files(F),
+	statistics(runtime,[_,_]),
+	set(ground_body,true),
+	set(depth_bound,1),
+	set(min_error,0.05),
+	format("~nGround body~n~n",[]),
+	test_filesi(F,ground_body(true)),
+	statistics(runtime,[_,T]),
+	T1 is T /1000,
+	format("Test successful, time ~f secs.~n",[T1]).
+t:-
+	format("Test unsuccessful.~n",[]).
+
+test_filesi([],_GB).
+
+test_filesi([H|T],GB):-
+        library_directory(LD),
+	atom_concat(LD,'/cplint/examples/',ExDir),
+	atom_concat(ExDir,H,NH),
+	p(NH),!,
+	findall(A,test(A,H,GB),L),
+	test_alli(H,L),
+	test_filesi(T,GB).
+
+test_alli(_F,[]).
+
+test_alli(F,[H|T]):-
+        copy_term(H,NH),
+	NH=(s(Q,_P),close_to('P',P)),!,
+	format("~a ~p.~n",[F,NH]),
+	si(Q,PL,PU,_Time),!,
+	format("Lower bound ~f, Upper bound ~f~n",[PL,PU]),
+	P>=PL-1e-7,P=<PU+1e-7,
+	test_alli(F,T).
+
+test_alli(F,[H|T]):-
+        copy_term(H,NH),
+	NH=(sc(Q,E,_P),close_to('P',P)),
+	format("~a ~p.~n",[F,NH]),
+	sci(Q,E,PL,PU,_Time),!,
+	format("Lower bound ~f, Upper bound ~f~n",[PL,PU]),
+	P>=PL-1e-10,P=<PU+1e-10,
+	test_alli(F,T).
+
+
+files([
+exapprox,exrange, 
+threesideddice,
+mendel,
+coin2,ex,trigger,throws,light]).
+
+test((s([death],P),close_to(P,0.305555555555556)),trigger,_).
+
+test((s([throws(mary),throws(john),break],P),close_to(P,0.46)),throws,_). 
+test((s([throws(mary),throws(john),\+break],P),close_to(P,0.04)),throws,_).
+test((s([\+ throws(mary),throws(john),break],P),close_to(P,0.3)),throws,_). 
+test((s([\+ throws(mary),throws(john),\+ break],P),close_to(P,0.2)),throws,_).
+
+test((s([push,replace],P),close_to(P,0.5)),light,_).
+test((s([push,light],P),close_to(P,0.5)),light,_).
+test((s([push,light,replace],P),close_to(P,0)),light,_).
+test((s([light,replace],P),close_to(P,0)),light,_).
+test((s([light],P),close_to(P,0.5)),light,_).
+test((s([replace],P),close_to(P,0.5)),light,_).
+
+
+test((s([\+ cites_cited(c1,p1)],P),close_to(P,0.7)),paper_ref_not,_).
+test((s([cites_citing(c1,p1)],P),close_to(P,0.14)),paper_ref_not,_).
+
+
+test((s([cites_cited(c1,p1)],P),close_to(P,0.181333333)),paper_ref,_).
+test((s([cites_cited(c1,p2)],P),close_to(P,0.181333333)),paper_ref,_).
+test((s([cites_cited(c1,p4)],P),close_to(P,0.181333333)),paper_ref,_).
+test((s([cites_cited(c1,p3)],P),close_to(P,0.228)),paper_ref,_).
+test((s([cites_cited(c1,p5)],P),close_to(P,0.228)),paper_ref,_).
+
+
+test((s([female(f)],P),close_to(P,0.6)),female,_).
+test((s([male(f)],P),close_to(P,0.4)),female,_).
+
+test((s([a],P),close_to(P,0.1719)),exapprox,ground_body(true)).
+test((s([a],P),close_to(P,0.099)),exapprox,ground_body(false)).
+
+test((s([a(1)],P),close_to(P,0.2775)),exrange,_).
+test((s([a(2)],P),close_to(P,0.36)),exrange,_).
+
+test((s([on(0,1)],P),close_to(P,0.333333333333333)),threesideddice,_).
+test((s([on(1,1)],P),close_to(P,0.222222222222222)),threesideddice,_).
+test((s([on(2,1)],P),close_to(P,0.148148147703704)),threesideddice,_).
+test((s([on(3,1)],P),close_to(P,0.0987654320987654)),threesideddice,_).  
+test((s([on(4,1)],P),close_to(P,0.0658436213991769)),threesideddice,_).
+
+test((sc([on(2,1)],[on(0,1)],P),close_to(P,0.222222222222222)),threesideddice,_).
+test((sc([on(2,1)],[on(1,1)],P),close_to(P,0.333333333333333)),threesideddice,_).
+test((sc([on(4,1)],[on(1,1)],P),close_to(P, 0.148148148148148)),threesideddice,_).
+test((sc([on(5,1)],[on(2,1)],P),close_to(P, 0.148148148148148)),threesideddice,_).
+
+
+
+test((s([cg(s,1,p)],P),close_to(P,0.75)),mendel,_).
+test((s([cg(s,1,w)],P),close_to(P,0.25)),mendel,_).
+test((s([cg(s,2,p)],P),close_to(P,0.25)),mendel,_).
+test((s([cg(s,2,w)],P),close_to(P,0.75)),mendel,_).
+test((s([cg(f,2,w)],P),close_to(P,0.5)),mendel,_).
+test((s([cg(s,2,w)],P),close_to(P,0.75)),mendel,_).
+
+test((s([a],P),close_to(P,0.226)),ex,_).
+
+test((s([heads(coin1)],P),close_to(P,0.51)),coin2,_).
+test((s([heads(coin2)],P),close_to(P,0.51)),coin2,_).
+
+test((s([tails(coin1)],P),close_to(P,0.49)),coin2,_).
+test((s([tails(coin2)],P),close_to(P,0.49)),coin2,_).
+
+
+
+test((s([student_rank(jane_doe,h)],P),close_to(P,0.465)),student,_).
+test((s([student_rank(jane_doe,l)],P),close_to(P,0.535)),student,_).
+
+test((s([course_rat(phil101,h)],P),close_to(P,0.330656)),student,_).
+test((s([course_rat(phil101,l)],P),close_to(P,0.669344)),student,_).
+
+
+test((s([professor_ability(p0,h)],P),close_to(P,0.5)),school,_).
+test((s([professor_ability(p0,m)],P),close_to(P,0.4)),school,_).
+test((s([professor_ability(p0,l)],P),close_to(P,0.1)),school,_).
+
+
+test((s([professor_popularity(p0,h)],P),close_to(P,0.531)),school,_).
+test((s([professor_popularity(p0,l)],P),close_to(P,0.175)),school,_).
+test((s([professor_popularity(p0,m)],P),close_to(P,0.294)),school,_).
+
+test((sc([professor_ability(p0,h)],[professor_popularity(p0,h)],P),close_to(P,0.847457627118644)),school,_).
+test((sc([professor_ability(p0,l)],[professor_popularity(p0,h)],P),close_to(P,0.00188323917137476)),school,_).
+test((sc([professor_ability(p0,m)],[professor_popularity(p0,h)],P),close_to(P,0.150659133709981)),school,_).
+
+test((sc([professor_popularity(p0,h)],[professor_ability(p0,h)],P),close_to(P,0.9)),school,_).
+test((sc([professor_popularity(p0,l)],[professor_ability(p0,h)],P),close_to(P,0.01)),school,_).
+test((sc([professor_popularity(p0,m)],[professor_ability(p0,h)],P),close_to(P,0.09)),school,_).
+
+test(( s([registration_grade(r0,1)],P),close_to(P,0.06675)),school,_).
+test(( s([registration_grade(r0,2)],P),close_to(P,0.16575)),school,_).
+test(( s([registration_grade(r0,3)],P),close_to(P, 0.356)),school,_).
+test(( s([registration_grade(r0,4)],P),close_to(P,0.4115)),school,_).
+
+test((sc([registration_grade(r0,1)],[registration_course(r0,C), course_difficulty(C,h)],P),close_to(P,0.15)),school,_).
+test((sc([registration_grade(r0,2)],[registration_course(r0,C), course_difficulty(C,h)],P),close_to(P,0.285)),school,_).
+test((sc([registration_grade(r0,3)],[registration_course(r0,C), course_difficulty(C,h)],P),close_to(P,0.424)),school,_).
+test((sc([registration_grade(r0,4)],[registration_course(r0,C), course_difficulty(C,h)],P),close_to(P,0.141)),school,_).
+
+test((sc([registration_grade(r0,1)], [registration_course(r0,C), course_difficulty(C,h), 
+	registration_student(r0,S), student_intelligence(S,h)],P),close_to(P,0.05)),school,_).
+test((sc([registration_grade(r0,2)], [registration_course(r0,C), course_difficulty(C,h), 
+	registration_student(r0,S), student_intelligence(S,h)],P),close_to(P,0.15)),school,_).
+test((sc([registration_grade(r0,3)], [registration_course(r0,C), course_difficulty(C,h), 
+	registration_student(r0,S), student_intelligence(S,h)],P),close_to(P,0.6)),school,_).
+test((sc([registration_grade(r0,4)], [registration_course(r0,C), course_difficulty(C,h), 
+	registration_student(r0,S), student_intelligence(S,h)],P),close_to(P,0.2)),school,_).
+
+test((sc([registration_grade(r0,1)],[registration_course(r0,C), course_difficulty(C,l), 
+	registration_student(r0,S), student_intelligence(S,h)],P),close_to(P,0.01)),school,_).
+test((sc([registration_grade(r0,2)],[registration_course(r0,C), course_difficulty(C,l), 
+	registration_student(r0,S), student_intelligence(S,h)],P),close_to(P,0.02)),school,_).
+test((sc([registration_grade(r0,3)],[registration_course(r0,C), course_difficulty(C,l), 
+	registration_student(r0,S), student_intelligence(S,h)],P),close_to(P,0.12)),school,_).
+test((sc([registration_grade(r0,4)],[registration_course(r0,C), course_difficulty(C,l), 
+	registration_student(r0,S), student_intelligence(S,h)],P),close_to(P,0.85)),school,_).
+
+test((s([registration_satisfaction(r0,1)],P),close_to(P,0.15197525)),school,_).
+test((s([registration_satisfaction(r0,2)],P),close_to(P,0.1533102)),school,_).
+test((s([registration_satisfaction(r0,3)],P),close_to(P,0.6947145)),school,_).
+
+test((sc([registration_satisfaction(r0,1)],[ registration_student(r0,S), 
+	student_intelligence(S,h)],P),close_to(P,0.0959225)),school,_).
+test((sc([registration_satisfaction(r0,2)],[ registration_student(r0,S), 
+	student_intelligence(S,h)],P),close_to(P,0.124515)),school,_).
+test((sc([registration_satisfaction(r0,3)],[ registration_student(r0,S), 
+	student_intelligence(S,h)],P),close_to(P,0.7795625)),school,_).
+
+test((sc([registration_satisfaction(r0,1)],[registration_grade(r0,4)],P),close_to(P,0.04)),school,_).
+test((sc([registration_satisfaction(r0,2)],[registration_grade(r0,4)],P),close_to(P,0.06)),school,_).
+test((sc([registration_satisfaction(r0,3)],[registration_grade(r0,4)],P),close_to(P,0.9)),school,_).
+
+test((sc([registration_satisfaction(r0,1)],[registration_grade(r0,1)],P),close_to(P,0.528)),school,_).
+test((sc([registration_satisfaction(r0,2)],[registration_grade(r0,1)],P),close_to(P,0.167)),school,_).
+test((sc([registration_satisfaction(r0,3)],[registration_grade(r0,1)],P),close_to(P,0.305)),school,_).
+
+test((sc([ registration_grade(r0,1)],[registration_satisfaction(r0,3)],P),close_to(P,0.0293052037923492)),school,_).
+test((sc([ registration_grade(r0,2)],[registration_satisfaction(r0,3)],P),close_to(P, 0.114760451955444)),school,_).
+test((sc([ registration_grade(r0,3)],[registration_satisfaction(r0,3)],P),close_to(P,0.322837654892765)),school,_).
+test((sc([ registration_grade(r0,4)],[registration_satisfaction(r0,3)],P),close_to(P,0.533096689359442)),school,_).
+
+test((s([course_rating(c0,h)],P),close_to(P,0.5392099)),school,_).
+test((s([course_rating(c0,l)],P),close_to(P, 0.2)),school,_).
+test((s([course_rating(c0,m)],P),close_to(P,0.2607901)),school,_).
+
+test((sc([course_difficulty(c0,h)],[course_rating(c0,h)],P),close_to(P,0.235185778302661)),school,_).
+test((sc([course_difficulty(c0,l)],[course_rating(c0,h)],P),close_to(P,0.259096503977393)),school,_).
+test((sc([course_difficulty(c0,m)],[course_rating(c0,h)],P),close_to(P,0.505717717719945)),school,_).
+
+test((s([course_difficulty(c0,h)],P),close_to(P,0.25)),school,_).
+test((s([course_difficulty(c0,l)],P),close_to(P,0.25)),school,_).
+test((s([course_difficulty(c0,m)],P),close_to(P,0.5)),school,_).
+
+
+test((s([student_ranking(s0,h)],P),close_to(P,0.6646250000000005)),school_simple,_).
+test((s([student_ranking(s0,l)],P),close_to(P,0.33537499999999987)),school_simple,_).
+