Fail to solve the transition path from arbitrary initial point to steady state

Perfect foresight simulations
1

I want to see the transition behavior from, say, 1/3 of the steady state, to the steady state. However, there is a following warning when find the solution for the linear system.

Warning: Matrix is close to singular or badly scaled.

Thus the Newton method fail to solve transition, and then it turns to homotopy method. However, it still fails with following warning.

Simulation of the perfect foresight model failed!Switching to a homotopy method…
Iter. | Lambda | status | Max. residual
++++++++++++++++++++++++++++++++++++++++++++++++++++++++
1 | 0.50000 | failed | 3.472947e+01
2 | 0.25000 | failed | 2.028938e+01
3 | 0.12500 | failed | 1.087585e+01
4 | 0.06250 | failed | 5.620716e+00
5 | 0.03125 | failed | 2.856055e+00
6 | 0.01562 | failed | 1.439452e+00
7 | 0.00781 | failed | 7.225821e-01
8 | 0.00391 | failed | 3.620051e-01
9 | 0.00195 | failed | 1.811810e-01
10 | 0.00098 | failed | 9.063515e-02
11 | 0.00049 | failed | 4.532873e-02
12 | 0.00024 | failed | 2.266716e-02
13 | 0.00012 | failed | 1.133427e-02
14 | 0.00006 | failed | 5.667312e-03
15 | 0.00003 | failed | 2.833699e-03
16 | 0.00002 | failed | 1.416861e-03
++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Failed to solve perfect foresight model

Total computing time : 0h00m12s
Note: warning(s) encountered in MATLAB/Octave code

2

Here I attach my mod file. Thanks a lot if anyone can give any diagnosis!!!

revise_GLC_steadystate_helper.m (1.5 KB)

revise_GLC.mod (2.5 KB)

3

Are you sure the timing in your model is correct?

Ku(+1)=(1-delta)*Ku+Iu;

looks inconsistent with Dynare’s standard timing.

4

Thanks a lot, Professor Jpfeifer!!! After modify this problem, the transition path can be solved.