What does this sentence mean? About SDP

@Mark_L_Stone, @Michal_Adamaszek

I am trying to solve a Stochastic MPC problem for a Network Controlled System (NCS) based control
I am using SDPT3 solver. My simulation run for the optimization problem gives out the following messages :

num. of constraints = 54
dim. of socp var = 27, num. of socp blk = 1
dim. of linear var = 59

SDPT3: Infeasible path-following algorithms

version predcorr gam expon scale_data
HKM 1 0.000 1 0
it pstep dstep pinfeas dinfeas gap prim-obj dual-obj cputime

0|0.000|0.000|1.0e+01|1.8e+00|1.3e+07|-5.500225e+06 0.000000e+00| 0:0:00| chol 1 1
1|0.948|0.740|5.3e-01|4.8e-01|3.3e+06| 2.963048e+04 -1.372817e+04| 0:0:00| chol 1 1
2|1.000|0.981|8.9e-08|9.2e-03|3.0e+05| 1.758180e+05 -2.369740e+04| 0:0:00| chol 1 1
3|0.884|0.936|7.0e-08|6.0e-04|3.8e+04| 1.549593e+04 -1.503332e+04| 0:0:00| chol 1 1
4|0.255|0.708|1.1e-07|1.8e-04|2.8e+04| 1.218268e+04 -1.153551e+04| 0:0:00| chol 1 1
5|0.680|0.815|3.4e-08|3.7e-05|1.5e+04| 2.350088e+03 -1.175631e+04| 0:0:00| chol 1 1
6|0.903|1.000|3.7e-09|3.7e-06|7.5e+03|-4.674813e+03 -1.197767e+04| 0:0:00| chol 1 1
7|1.000|1.000|4.3e-10|1.1e-06|2.7e+03|-8.352582e+03 -1.103458e+04| 0:0:00| chol 1 1
8|0.852|1.000|2.7e-10|3.3e-07|1.2e+03|-9.390996e+03 -1.055545e+04| 0:0:00| chol 1 1
9|1.000|1.000|7.7e-11|9.9e-08|3.5e+02|-9.942374e+03 -1.028593e+04| 0:0:00| chol 1 1
10|0.863|1.000|1.2e-11|3.0e-08|1.3e+02|-1.004625e+04 -1.017981e+04| 0:0:00| chol 1 1
11|1.000|0.966|1.2e-12|9.6e-09|2.4e+01|-1.010277e+04 -1.012632e+04| 0:0:00| chol 1 1
12|0.938|1.000|2.1e-12|2.7e-09|6.2e+00|-1.010845e+04 -1.011462e+04| 0:0:00| chol 1 1
13|0.988|0.942|2.5e-12|9.1e-10|4.7e-01|-1.011038e+04 -1.011083e+04| 0:0:00| chol 1 1
14|0.974|0.949|1.8e-11|4.7e-11|2.4e-02|-1.011046e+04 -1.011048e+04| 0:0:00| chol 1 1
15|0.971|0.959|2.8e-12|3.4e-12|1.0e-03|-1.011046e+04 -1.011046e+04| 0:0:00|
stop: max(relative gap, infeasibilities) < 1.00e-07

number of iterations = 15
primal objective value = -1.01104629e+04
dual objective value = -1.01104638e+04
gap := trace(XZ) = 9.96e-04
relative gap = 4.92e-08
actual relative gap = 4.63e-08
rel. primal infeas = 2.77e-12
rel. dual infeas = 3.42e-12
norm(X), norm(y), norm(Z) = 3.3e+03, 1.8e+04, 5.5e+03
norm(A), norm(b), norm© = 5.1e+01, 2.3e+03, 2.0e+04
Total CPU time (secs) = 0.06
CPU time per iteration = 0.00
termination code = 0
DIMACS errors: 4.2e-12 0.0e+00 4.8e-12 0.0e+00 4.6e-08 4.9e-08

Percentage of CPU time spent in various parts

preproc Xchol Zchol pred pred_steplen corr corr_steplen misc
4.3 4.3 4.3 42.0 5.8 5.8 11.6 2.9 2.9 15.9

The results of my simulation are not as expected. I want to know what ‘DIMACS errors’ stand for?
Also could you please let me know, based upon these messages, is there anything wrong with my simulation?

It is a convex quadratic program. Additive process noise is gaussian with zero mean and variance I3, i.e. mu = 0, variance = eye(3) and standard_deviation = sqrt(variance). Channel dropout effect is considered to b binary Bernoulli.
The optimization problem to be solved is as below:

I suggest you ask Johan for further assistance on your existing thread https://groups.google.com/g/yalmip/c/3TBLhZ73tEk

I have posed question regarding DIMACS errors to Johan. He shared a link to some old article which I could not understand… :thinking:

However, he is very responsive and supportive…

In my experience, there may exist some mistakes in your code. If you can make sure that the formulated problem is feasible, please check your code.

All the DIMACS errors are small so it seems that the optimizer solves your problem quite well.
In other words the software does what it is supposed to do. This implies your expectations and/or the model is wrong.

1 Like

Solution is feasible…