////////////////////////////////////////////////////////////////////////////////////////////////////
// FILTER SQP SOLVER STARTED //
////////////////////////////////////////////////////////////////////////////////////////////////////
=== OUTER ITERATION 0 STARTED ==================================================================
trustRad = 1.000e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 0.000e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 0.000e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=0.00e+00
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-5.89e+03, actual=-5.95e+03, ratio= 1.01
>> change in the constraints violation: predicted=0.00e+00, actual=0.00e+00, ratio=-nan(ind)
>> target and constraints are:
targetF: -3.370041606e+05 -> -3.429507004e+05 (delta= -5.947e+03)
sum-of-constr: 0.000000000e+00 -> 0.000000000e+00 (delta= 0.000e+00)
> analyzing Hessian update
>> applying conservative curvature decrease to correct too low (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=2.00e-01
>> printing information about Hessian:
mindiag(Bk) = 2.000e-01
maxdiag(Bk) = 2.000e-01
> the step was successfully performed
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 1.000 (ratio of step length to trust radius)
newTrustRad = 1.410e-01, trust radius increased
=== OUTER ITERATION 1 STARTED ==================================================================
trustRad = 1.410e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 0.000e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.000e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=0.00e+00
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-8.47e+03, actual=-8.58e+03, ratio= 1.01
>> change in the constraints violation: predicted=0.00e+00, actual=0.00e+00, ratio=-nan(ind)
>> target and constraints are:
targetF: -3.429507004e+05 -> -3.515328324e+05 (delta= -8.582e+03)
sum-of-constr: 0.000000000e+00 -> 0.000000000e+00 (delta= 0.000e+00)
> analyzing Hessian update
>> applying conservative curvature decrease to correct too low (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=4.00e-02
>> printing information about Hessian:
mindiag(Bk) = 4.000e-02
maxdiag(Bk) = 4.000e-02
> the step was successfully performed
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 1.000 (ratio of step length to trust radius)
newTrustRad = 3.976e-01, trust radius increased
=== OUTER ITERATION 2 STARTED ==================================================================
trustRad = 3.976e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 0.000e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 2.410e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=0.00e+00
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-2.45e+04, actual=-2.54e+04, ratio= 1.04
>> change in the constraints violation: predicted=0.00e+00, actual=4.47e-01, ratio= inf
>> target and constraints are:
targetF: -3.515328324e+05 -> -3.769792622e+05 (delta= -2.545e+04)
sum-of-constr: 0.000000000e+00 -> 4.470129677e-01 (delta= 4.470e-01)
> analyzing Hessian update
>> applying conservative curvature decrease to correct too low (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=8.00e-03
>> printing information about Hessian:
mindiag(Bk) = 8.000e-03
maxdiag(Bk) = 8.000e-03
> the step was successfully performed
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 1.000 (ratio of step length to trust radius)
newTrustRad = 2.243e+00, trust radius increased
=== OUTER ITERATION 3 STARTED ==================================================================
trustRad = 2.243e+00
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 2.235e-01 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 6.386e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.06e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-9.53e+04, actual=-1.12e+05, ratio= 1.17
>> change in the constraints violation: predicted=-4.47e-01, actual=1.97e+01, ratio=-44.14
>> target and constraints are:
targetF: -3.769792622e+05 -> -4.890003774e+05 (delta= -1.120e+05)
sum-of-constr: 4.470129677e-01 -> 2.017784673e+01 (delta= 1.973e+01)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=4.92e-02
>> printing information about Hessian:
mindiag(Bk) = 3.241e-02
maxdiag(Bk) = 5.839e-02
> the candidate point is not accepted by the filter, decreasing trust radius
> restarting iteration
=== OUTER ITERATION 4 STARTED ==================================================================
trustRad = 1.121e+00
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 2.235e-01 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 6.386e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.06e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-4.62e+04, actual=-5.03e+04, ratio= 1.09
>> change in the constraints violation: predicted=-4.47e-01, actual=4.64e+00, ratio=-10.39
>> target and constraints are:
targetF: -3.769792622e+05 -> -4.272613986e+05 (delta= -5.028e+04)
sum-of-constr: 4.470129677e-01 -> 5.090394942e+00 (delta= 4.643e+00)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=2.66e-01
>> printing information about Hessian:
mindiag(Bk) = 1.616e-01
maxdiag(Bk) = 3.252e-01
> the step was successfully performed
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 1.000 (ratio of step length to trust radius)
newTrustRad = 1.581e+00, trust radius increased
=== OUTER ITERATION 5 STARTED ==================================================================
trustRad = 1.581e+00
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 2.545e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.760e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.23e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=5.79e+03, actual=3.76e+03, ratio= 0.65
>> change in the constraints violation: predicted=-5.09e+00, actual=5.90e+00, ratio=-1.16
>> target and constraints are:
targetF: -4.272613986e+05 -> -4.235059380e+05 (delta= 3.755e+03)
sum-of-constr: 5.090394942e+00 -> 1.099532724e+01 (delta= 5.905e+00)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=1.10e+00
>> printing information about Hessian:
mindiag(Bk) = 5.832e-01
maxdiag(Bk) = 1.320e+00
> the candidate point is not accepted by the filter, decreasing trust radius
> restarting iteration
=== OUTER ITERATION 6 STARTED ==================================================================
trustRad = 7.905e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 2.545e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.760e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.23e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=2.24e+04, actual=2.10e+04, ratio= 0.94
>> change in the constraints violation: predicted=-5.09e+00, actual=-1.57e+00, ratio= 0.31
>> target and constraints are:
targetF: -4.272613986e+05 -> -4.062869873e+05 (delta= 2.097e+04)
sum-of-constr: 5.090394942e+00 -> 3.520784691e+00 (delta= -1.570e+00)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=5.39e+00
>> printing information about Hessian:
mindiag(Bk) = 2.818e+00
maxdiag(Bk) = 5.310e+00
> the step was successfully performed
>> the initial point was added to the filter (1 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 1.000 (ratio of step length to trust radius)
newTrustRad = 1.115e+00, trust radius increased
=== OUTER ITERATION 7 STARTED ==================================================================
trustRad = 1.115e+00
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 1.760e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.254e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.90e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-3.04e+04, actual=-3.40e+04, ratio= 1.12
>> change in the constraints violation: predicted=-3.52e+00, actual=5.58e+00, ratio=-1.58
>> target and constraints are:
targetF: -4.062869873e+05 -> -4.402916069e+05 (delta= -3.400e+04)
sum-of-constr: 3.520784691e+00 -> 9.099417037e+00 (delta= 5.579e+00)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=2.63e+01
>> printing information about Hessian:
mindiag(Bk) = 1.371e+01
maxdiag(Bk) = 2.424e+01
> the step was successfully performed
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 1.000 (ratio of step length to trust radius)
newTrustRad = 3.143e+00, trust radius increased
=== OUTER ITERATION 8 STARTED ==================================================================
trustRad = 3.143e+00
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 4.550e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 2.209e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.03e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-8.17e+04, actual=-9.19e+04, ratio= 1.13
>> change in the constraints violation: predicted=-9.10e+00, actual=3.06e+01, ratio=-3.37
>> target and constraints are:
targetF: -4.402916069e+05 -> -5.322174838e+05 (delta= -9.193e+04)
sum-of-constr: 9.099417037e+00 -> 3.972764262e+01 (delta= 3.063e+01)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=9.89e+01
>> printing information about Hessian:
mindiag(Bk) = 5.282e+01
maxdiag(Bk) = 9.952e+01
> the candidate point is not accepted by the filter, decreasing trust radius
> restarting iteration
=== OUTER ITERATION 9 STARTED ==================================================================
trustRad = 1.572e+00
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 4.550e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 2.209e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.03e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-1.17e+04, actual=-1.37e+04, ratio= 1.17
>> change in the constraints violation: predicted=-9.10e+00, actual=1.03e+00, ratio=-0.11
>> target and constraints are:
targetF: -4.402916069e+05 -> -4.539557687e+05 (delta= -1.366e+04)
sum-of-constr: 9.099417037e+00 -> 1.012990852e+01 (delta= 1.030e+00)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=4.97e+02
>> printing information about Hessian:
mindiag(Bk) = 2.534e+02
maxdiag(Bk) = 4.893e+02
> the candidate point is not accepted by the filter, decreasing trust radius
> restarting iteration
=== OUTER ITERATION 10 STARTED ==================================================================
trustRad = 7.858e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 4.550e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 2.209e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.03e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=2.34e+04, actual=2.15e+04, ratio= 0.92
>> change in the constraints violation: predicted=-9.10e+00, actual=-5.67e+00, ratio= 0.62
>> target and constraints are:
targetF: -4.402916069e+05 -> -4.187466287e+05 (delta= 2.154e+04)
sum-of-constr: 9.099417037e+00 -> 3.427108645e+00 (delta= -5.672e+00)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=2.11e+03
>> printing information about Hessian:
mindiag(Bk) = 1.084e+03
maxdiag(Bk) = 2.005e+03
> the step was successfully performed
>> the initial point was added to the filter (2 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 1.000 (ratio of step length to trust radius)
newTrustRad = 1.108e+00, trust radius increased
=== OUTER ITERATION 11 STARTED ==================================================================
trustRad = 1.108e+00
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 1.714e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.517e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.38e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=2.16e+04, actual=2.05e+04, ratio= 0.95
>> change in the constraints violation: predicted=-3.43e+00, actual=1.27e+00, ratio=-0.37
>> target and constraints are:
targetF: -4.187466287e+05 -> -3.982547560e+05 (delta= 2.049e+04)
sum-of-constr: 3.427108645e+00 -> 4.693744847e+00 (delta= 1.267e+00)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=2.10e+03
>> printing information about Hessian:
mindiag(Bk) = 1.086e+03
maxdiag(Bk) = 2.855e+03
> the step was successfully performed
>> the initial point was added to the filter (3 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 1.000 (ratio of step length to trust radius)
newTrustRad = 3.125e+00, trust radius increased
=== OUTER ITERATION 12 STARTED ==================================================================
trustRad = 3.125e+00
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 2.462e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.436e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.36e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-1.44e+05, actual=-2.18e+05, ratio= 1.51
>> change in the constraints violation: predicted=-4.69e+00, actual=5.23e+01, ratio=-11.15
>> target and constraints are:
targetF: -3.982547560e+05 -> -6.165182898e+05 (delta= -2.183e+05)
sum-of-constr: 4.693744847e+00 -> 5.704299036e+01 (delta= 5.235e+01)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=9.15e+03
>> printing information about Hessian:
mindiag(Bk) = 4.709e+03
maxdiag(Bk) = 9.323e+03
> the candidate point is not accepted by the filter, decreasing trust radius
> restarting iteration
=== OUTER ITERATION 13 STARTED ==================================================================
trustRad = 1.562e+00
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 2.462e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.436e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.07e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-2.76e+04, actual=-7.34e+04, ratio= 2.66
>> change in the constraints violation: predicted=-4.69e+00, actual=6.84e+00, ratio=-1.46
>> target and constraints are:
targetF: -3.982547560e+05 -> -4.716935719e+05 (delta= -7.344e+04)
sum-of-constr: 4.693744847e+00 -> 1.153817457e+01 (delta= 6.844e+00)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=1.77e+04
>> printing information about Hessian:
mindiag(Bk) = 9.421e+03
maxdiag(Bk) = 1.774e+04
> the candidate point is not accepted by the filter, decreasing trust radius
> restarting iteration
=== OUTER ITERATION 14 STARTED ==================================================================
trustRad = 7.811e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 2.462e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.436e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.25e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=3.20e+03, actual=-1.33e+04, ratio=-4.16
>> change in the constraints violation: predicted=-4.69e+00, actual=-2.08e+00, ratio= 0.44
>> target and constraints are:
targetF: -3.982547560e+05 -> -4.115367277e+05 (delta= -1.328e+04)
sum-of-constr: 4.693744847e+00 -> 2.617162241e+00 (delta= -2.077e+00)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.13e+04
>> printing information about Hessian:
mindiag(Bk) = 1.138e+04
maxdiag(Bk) = 2.245e+04
> the step was successfully performed
>> the initial point was added to the filter (4 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 1.000 (ratio of step length to trust radius)
newTrustRad = 1.101e+00, trust radius increased
=== OUTER ITERATION 15 STARTED ==================================================================
trustRad = 1.101e+00
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 1.333e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.475e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.43e+04
> proposed step with relative len 0.296759 (compared to trust radius)
>> change in the target function: predicted=2.28e+04, actual=2.06e+04, ratio= 0.91
>> change in the constraints violation: predicted=-2.62e+00, actual=-2.08e+00, ratio= 0.80
>> target and constraints are:
targetF: -4.115367277e+05 -> -3.908891339e+05 (delta= 2.065e+04)
sum-of-constr: 2.617162241e+00 -> 5.324067509e-01 (delta= -2.085e+00)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.07e+04
>> printing information about Hessian:
mindiag(Bk) = 1.138e+04
maxdiag(Bk) = 2.316e+04
> the step was successfully performed
>> the initial point was added to the filter (4 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.297 (ratio of step length to trust radius)
newTrustRad = 6.537e-01, trust radius decreased
=== OUTER ITERATION 16 STARTED ==================================================================
trustRad = 6.537e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 2.812e-01 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.025e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.83e+04
> proposed step with relative len 0.902679 (compared to trust radius)
>> change in the target function: predicted=4.56e+03, actual=2.66e+03, ratio= 0.58
>> change in the constraints violation: predicted=-5.32e-01, actual=7.43e-01, ratio=-1.39
>> target and constraints are:
targetF: -3.908891339e+05 -> -3.882263296e+05 (delta= 2.663e+03)
sum-of-constr: 5.324067509e-01 -> 1.274976973e+00 (delta= 7.426e-01)
> analyzing Hessian update
>> applying damping to correct too high (or rapidly changing) curvature
>> LBFGS update applied, diagonal scaling sigma=1.83e+04
>> printing information about Hessian:
mindiag(Bk) = 1.042e+04
maxdiag(Bk) = 2.784e+04
> the step was successfully performed
>> the initial point was added to the filter (5 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.903 (ratio of step length to trust radius)
newTrustRad = 6.537e-01
=== OUTER ITERATION 17 STARTED ==================================================================
trustRad = 6.537e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 6.565e-01 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.234e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.61e+04
> proposed step with relative len 1.000000 (compared to trust radius)
>> change in the target function: predicted=-3.31e+03, actual=-1.50e+04, ratio= 4.54
>> change in the constraints violation: predicted=-1.27e+00, actual=4.45e-01, ratio=-0.35
>> target and constraints are:
targetF: -3.882263296e+05 -> -4.032448171e+05 (delta= -1.502e+04)
sum-of-constr: 1.274976973e+00 -> 1.719821802e+00 (delta= 4.448e-01)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.70e+04
>> printing information about Hessian:
mindiag(Bk) = 1.498e+04
maxdiag(Bk) = 3.965e+04
> the step was successfully performed
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 1.000 (ratio of step length to trust radius)
newTrustRad = 9.217e-01, trust radius increased
=== OUTER ITERATION 18 STARTED ==================================================================
trustRad = 9.217e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 8.599e-01 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.244e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.80e+04
> proposed step with relative len 0.317027 (compared to trust radius)
>> change in the target function: predicted=1.79e+04, actual=1.55e+04, ratio= 0.87
>> change in the constraints violation: predicted=-1.72e+00, actual=-1.47e+00, ratio= 0.85
>> target and constraints are:
targetF: -4.032448171e+05 -> -3.877076856e+05 (delta= 1.554e+04)
sum-of-constr: 1.719821802e+00 -> 2.501614917e-01 (delta= -1.470e+00)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.69e+04
>> printing information about Hessian:
mindiag(Bk) = 1.486e+04
maxdiag(Bk) = 4.119e+04
> the step was successfully performed
>> the initial point was added to the filter (6 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.317 (ratio of step length to trust radius)
newTrustRad = 5.844e-01, trust radius decreased
=== OUTER ITERATION 19 STARTED ==================================================================
trustRad = 5.844e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 1.300e-01 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 1.008e+00
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=1.99e+04
> proposed step with relative len 0.109997 (compared to trust radius)
>> change in the target function: predicted=3.02e+03, actual=2.91e+03, ratio= 0.96
>> change in the constraints violation: predicted=-2.50e-01, actual=-2.41e-01, ratio= 0.96
>> target and constraints are:
targetF: -3.877076856e+05 -> -3.847993320e+05 (delta= 2.908e+03)
sum-of-constr: 2.501614917e-01 -> 9.595839740e-03 (delta= -2.406e-01)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.77e+04
>> printing information about Hessian:
mindiag(Bk) = 1.324e+04
maxdiag(Bk) = 4.198e+04
> the step was successfully performed
>> the initial point was added to the filter (7 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.110 (ratio of step length to trust radius)
newTrustRad = 1.286e-01, trust radius decreased
=== OUTER ITERATION 20 STARTED ==================================================================
trustRad = 1.286e-01
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 5.116e-03 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 9.435e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=2.00e+04
> proposed step with relative len 0.217968 (compared to trust radius)
>> change in the target function: predicted=1.08e+02, actual=9.74e+01, ratio= 0.90
>> change in the constraints violation: predicted=-9.60e-03, actual=-8.45e-03, ratio= 0.88
>> target and constraints are:
targetF: -3.847993320e+05 -> -3.847018963e+05 (delta= 9.744e+01)
sum-of-constr: 9.595839740e-03 -> 1.143609774e-03 (delta= -8.452e-03)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.08e+04
>> printing information about Hessian:
mindiag(Bk) = 1.063e+04
maxdiag(Bk) = 4.001e+04
> the step was successfully performed
>> the initial point was added to the filter (8 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.218 (ratio of step length to trust radius)
newTrustRad = 5.605e-02, trust radius decreased
=== OUTER ITERATION 21 STARTED ==================================================================
trustRad = 5.605e-02
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 9.960e-04 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 9.445e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=2.01e+04
> proposed step with relative len 0.201623 (compared to trust radius)
>> change in the target function: predicted=7.39e+00, actual=6.29e+00, ratio= 0.85
>> change in the constraints violation: predicted=-1.14e-03, actual=-9.95e-04, ratio= 0.87
>> target and constraints are:
targetF: -3.847018963e+05 -> -3.846956028e+05 (delta= 6.293e+00)
sum-of-constr: 1.143609774e-03 -> 1.488323456e-04 (delta= -9.948e-04)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.11e+04
>> printing information about Hessian:
mindiag(Bk) = 9.754e+03
maxdiag(Bk) = 3.948e+04
> the step was successfully performed
>> the initial point was added to the filter (9 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.202 (ratio of step length to trust radius)
newTrustRad = 2.260e-02, trust radius decreased
=== OUTER ITERATION 22 STARTED ==================================================================
trustRad = 2.260e-02
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 1.300e-04 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 9.440e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=2.01e+04
> proposed step with relative len 0.041817 (compared to trust radius)
>> change in the target function: predicted=1.09e+00, actual=1.08e+00, ratio= 0.99
>> change in the constraints violation: predicted=-1.49e-04, actual=-1.47e-04, ratio= 0.99
>> target and constraints are:
targetF: -3.846956028e+05 -> -3.846945278e+05 (delta= 1.075e+00)
sum-of-constr: 1.488323456e-04 -> 1.333944441e-06 (delta= -1.475e-04)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.97e+04
>> printing information about Hessian:
mindiag(Bk) = 1.070e+04
maxdiag(Bk) = 4.490e+04
> the step was successfully performed
>> the initial point was added to the filter (10 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.042 (ratio of step length to trust radius)
newTrustRad = 2.260e-03, trust radius decreased
=== OUTER ITERATION 23 STARTED ==================================================================
trustRad = 2.260e-03
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 9.038e-07 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 9.440e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=2.01e+04
> proposed step with relative len 0.109906 (compared to trust radius)
>> change in the target function: predicted=1.23e-02, actual=1.10e-02, ratio= 0.89
>> change in the constraints violation: predicted=-1.33e-06, actual=-1.21e-06, ratio= 0.91
>> target and constraints are:
targetF: -3.846945278e+05 -> -3.846945167e+05 (delta= 1.102e-02)
sum-of-constr: 1.333944441e-06 -> 1.242390859e-07 (delta= -1.210e-06)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.34e+04
>> printing information about Hessian:
mindiag(Bk) = 1.085e+04
maxdiag(Bk) = 4.215e+04
> the step was successfully performed
>> the initial point was added to the filter (11 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.110 (ratio of step length to trust radius)
newTrustRad = 4.968e-04, trust radius decreased
=== OUTER ITERATION 24 STARTED ==================================================================
trustRad = 4.968e-04
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 1.032e-07 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 9.440e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=2.01e+04
> proposed step with relative len 0.331691 (compared to trust radius)
>> change in the target function: predicted=7.72e-04, actual=5.43e-04, ratio= 0.70
>> change in the constraints violation: predicted=-1.24e-07, actual=-9.14e-08, ratio= 0.74
>> target and constraints are:
targetF: -3.846945167e+05 -> -3.846945162e+05 (delta= 5.434e-04)
sum-of-constr: 1.242390859e-07 -> 3.288875616e-08 (delta= -9.135e-08)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=1.64e+04
>> printing information about Hessian:
mindiag(Bk) = 1.039e+04
maxdiag(Bk) = 4.109e+04
> the step was successfully performed
>> the initial point was added to the filter (12 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.332 (ratio of step length to trust radius)
newTrustRad = 3.296e-04, trust radius decreased
=== OUTER ITERATION 25 STARTED ==================================================================
trustRad = 3.296e-04
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 3.061e-08 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 9.440e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=2.01e+04
> proposed step with relative len 0.040986 (compared to trust radius)
>> change in the target function: predicted=2.13e-04, actual=2.11e-04, ratio= 0.99
>> change in the constraints violation: predicted=-3.29e-08, actual=-3.26e-08, ratio= 0.99
>> target and constraints are:
targetF: -3.846945162e+05 -> -3.846945160e+05 (delta= 2.108e-04)
sum-of-constr: 3.288875616e-08 -> 2.744056093e-10 (delta= -3.261e-08)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.05e+04
>> printing information about Hessian:
mindiag(Bk) = 1.057e+04
maxdiag(Bk) = 4.046e+04
> the step was successfully performed
>> the initial point was added to the filter (13 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.041 (ratio of step length to trust radius)
newTrustRad = 3.296e-05, trust radius decreased
=== OUTER ITERATION 26 STARTED ==================================================================
trustRad = 3.296e-05
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 2.415e-10 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 9.440e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=2.01e+04
> proposed step with relative len 0.010046 (compared to trust radius)
>> change in the target function: predicted=1.99e-06, actual=1.99e-06, ratio= 1.00
>> change in the constraints violation: predicted=-2.74e-10, actual=-2.74e-10, ratio= 1.00
>> target and constraints are:
targetF: -3.846945160e+05 -> -3.846945160e+05 (delta= 1.988e-06)
sum-of-constr: 2.744056093e-10 -> 2.986499936e-13 (delta= -2.741e-10)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=3.34e+04
>> printing information about Hessian:
mindiag(Bk) = 1.075e+04
maxdiag(Bk) = 3.921e+04
> the step was successfully performed
>> the initial point was added to the filter (14 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.010 (ratio of step length to trust radius)
newTrustRad = 3.296e-06, trust radius decreased
=== OUTER ITERATION 27 STARTED ==================================================================
trustRad = 3.296e-06
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 1.528e-13 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 9.440e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=2.01e+04
> proposed step with relative len 0.018132 (compared to trust radius)
>> change in the target function: predicted=3.83e-09, actual=3.73e-09, ratio= 0.97
>> change in the constraints violation: predicted=-2.99e-13, actual=-2.99e-13, ratio= 1.00
>> target and constraints are:
targetF: -3.846945160e+05 -> -3.846945160e+05 (delta= 3.725e-09)
sum-of-constr: 2.986499936e-13 -> 0.000000000e+00 (delta= -2.986e-13)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=2.74e+04
>> printing information about Hessian:
mindiag(Bk) = 1.291e+04
maxdiag(Bk) = 4.137e+04
> the step was successfully performed
>> the initial point was added to the filter (15 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.018 (ratio of step length to trust radius)
newTrustRad = 3.296e-07, trust radius decreased
=== OUTER ITERATION 28 STARTED ==================================================================
trustRad = 3.296e-07
lin.violation = 0.000e+00 (scaled violation of linear constraints)
nlc.violation = 0.000e+00 (scaled violation of nonlinear constraints)
gamma0 = 0.000e+00 (Hessian 2-norm estimate for target)
gammaMax = 0.000e+00 (maximum over Hessian 2-norm estimates for target/constraints)
arg(gammaMax) = 0 (function index; 0 for target, >0 for nonlinear constraints)
|x|inf = 9.440e-01
--- filter SQP step --------------------------------------------------------------------------------
> the QP subproblem was solved with TerminationType=1, max|lagBoxMult|=0.00e+00, max|lagNonBoxMult|=2.01e+04
> proposed step with relative len 0.013059 (compared to trust radius)
>> change in the target function: predicted=3.20e-12, actual=0.00e+00, ratio= 0.00
>> change in the constraints violation: predicted=0.00e+00, actual=0.00e+00, ratio=-nan(ind)
>> target and constraints are:
targetF: -3.846945160e+05 -> -3.846945160e+05 (delta= 0.000e+00)
sum-of-constr: 0.000000000e+00 -> 0.000000000e+00 (delta= 0.000e+00)
> analyzing Hessian update
>> LBFGS update applied, diagonal scaling sigma=3.17e+04
>> printing information about Hessian:
mindiag(Bk) = 1.330e+04
maxdiag(Bk) = 4.460e+04
> the step was successfully performed
>> the initial point was added to the filter (16 entries in the filter)
--- iteration ends ---------------------------------------------------------------------------------
deltaMax = 0.013 (ratio of step length to trust radius)
newTrustRad = 3.296e-08, trust radius decreased
> stopping condition met: relative change in target is less than 0.000e+00
=== STOPPED ========================================================================================
raw target: -3.846945159733e+05
total time: 0.0 ms
* incl cbck: 0.0 ms
* incl QP: 0.0 ms
////////////////////////////////////////////////////////////////////////////////////////////////////
// OPTGUARD INTEGRITY CHECKER REPORT //
////////////////////////////////////////////////////////////////////////////////////////////////////
> no discontinuity/nonsmoothness/bad-gradient suspicions were raised during optimization