%The formula after virtual sub, has 3 forall quantifiers left; %qepcad runs forever; %slfq can simplify the formula, IF we remove the last two conjuncts. % We get the following...it does not have the FORALL variables!! % [ c >= 0 /\ b >= 0 /\ [ 2000 c b - b + 2000 c = 0 \/ [ 2200 c - 1 > 0 /\ b > 0 ] \/ [ 2200 c - 1 < 0 /\ b > 0 ] ] ] % There were 3174 QEPCADB calls! % System time: 3320 milliseconds. % System time after the initialization: 2800 milliseconds. % c = 1/2000, b = anything! or c=1/2200 and b=10; % subbing c=1/2200 and b=10 yield TRUE; hence verified. % subbing c=1/2000 and b left as param gives big formula; not tried simplifying on echo; load_package redlog; rlset r; load_package qepcad; on rlqefbqepcad; % use QEPCAD as the fallback QE rlqepcadn(100*10^6); % The number of cells used by QEPCAD. The % default is (only) 1 * 10^6. rlqepcadl(200*10^3); % The size of the prime list used by QEPCAD. The % default is (only) 2 * 10^3. on rlverbose; on time; off rlqefb; % No fallback QE phi1 := ( beta^2 <= 1 impl ( 2*(we + b*intwe)*(b*we -5*we-10*intwe+(L-optL)*beta) + 2*a*intwe*we + 2*c*(L-optL)*(-1000*intwe-2200*we)*beta <= 0)); phi3 := ( beta^2 <= 1 impl ( 2*(we + (10/22)*intwe)*((10/22)*we -5*we-10*intwe+(L-optL)*beta) + 2*a*intwe*we + 2*(1/2200)*(L-optL)*(-1000*intwe-2200*we)*beta <= 0)); %we get c=1/2200 and b = 1000*c % we also get 14641*a^2 - 353320*a + 921600 <= 0 % which wolfram tells me is 360/121 <= a <= 2560/121...i.e. 3 <= a <= 21 %phi := phi1 and phi2 and phi3; % we get (we+ b*intwe)^2 + a*intwe^2 + c*(L-Lopt)^2 as a Lyapunov.... phi := phi1; phi := all({intwe, we, L, optL, beta}, phi ) ; psi := rlqe( phi ); %psi1 := second psi; psi1 := psi; psi2 := rlqepcad( rlex(psi1), "adaptive-final.qepcad" ); %psi2 := rlqepcad( psi ); %psi3 := rlqepcad( rlex(psi2), "adaptive2-final.qepcad" ); off nat; % Output same syntax as input (for reduce) out "adaptive-final.out"; psi; %psi3; shut "adaptive-final.out"; end;