[Couenne-tickets] [Couenne, a solver for non-convex MINLP problems] #14: feasible model declared infeasible

Couenne, a solver for non-convex MINLP problems coin-trac at coin-or.org
Wed Sep 28 21:00:59 EDT 2011


#14: feasible model declared infeasible
-------------------------+--------------------------------------------------
  Reporter:  stefan      |       Owner:  somebody
      Type:  defect      |      Status:  new     
  Priority:  major       |   Milestone:          
 Component:  component1  |     Version:          
Resolution:              |    Keywords:          
-------------------------+--------------------------------------------------

Comment (by stefan):

 Setting {{{use_semiaux no}}} does not help.

 I increased some print levels and below is the output.
 It seems that probing finds the problem infeasible.
 Can you see, why?

 {{{
 objectives:
 min ((x_0^2)+1-2*x_0)
 constraints:
 (-x_0+3*x_1+x_2-x_9) = -1
 (-x_0+x_1+3*x_2-x_10) = 2
 (-x_0+3*x_3+x_4+x_11) = 4
 (-x_0+x_3+3*x_4+x_12) = 10
 (-x_0+x_1+x_2+x_3+x_4+10*x_5+x_6+x_7+x_8) = 16
 (-x_0+3*x_6+x_7+x_8-x_13+x_14) = 0
 (-x_0+x_6+3*x_7+x_8-x_15+x_16) = 4
 (-x_0+x_6+x_7+3*x_8-x_17+x_18) = 10
 (x_1*x_9) = 0
 (x_2*x_10) = 0
 ((-(x_3*x_11))+2*x_11) = 0
 ((-(x_4*x_12))+2*x_12) = 0
 (x_6*x_13) = 0
 ((-(x_6*x_14))+2*x_14) = 0
 (x_7*x_15) = 0
 ((-(x_7*x_16))+2*x_16) = 0
 (x_8*x_17) = 0
 ((-(x_8*x_18))+2*x_18) = 0
 variables:
 x_0 [ -1e+50 , 1e+50 ]
 x_1 [ 0 , 1e+50 ]
 x_2 [ 0 , 1e+50 ]
 x_3 [ -1e+50 , 2 ]
 x_4 [ -1e+50 , 2 ]
 x_5 [ -1e+50 , 1e+50 ]
 x_6 [ 0 , 2 ]
 x_7 [ 0 , 2 ]
 x_8 [ 0 , 2 ]
 x_9 [ 0 , 1e+50 ]
 x_10 [ 0 , 1e+50 ]
 x_11 [ 0 , 1e+50 ]
 x_12 [ 0 , 1e+50 ]
 x_13 [ 0 , 1e+50 ]
 x_14 [ 0 , 1e+50 ]
 x_15 [ 0 , 1e+50 ]
 x_16 [ 0 , 1e+50 ]
 x_17 [ 0 , 1e+50 ]
 x_18 [ 0 , 1e+50 ]
 end
 Problem size before reformulation: 19 variables (0 integer), 18
 constraints.
 Couenne: initial solution (value 0) is MINLP feasible
 Couenne: new MINLP solution, value 0.0000000000e+00
 created Expression Object: w_19 := (x_0^2) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Expression Object: w_21 := (x_1*w_9) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Expression Object: w_23 := (x_2*w_10) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Expression Object: w_25 := (x_3*w_11) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Expression Object: w_27 := (x_4*w_12) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Expression Object: w_29 := (x_6*x_13) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Expression Object: w_31 := (x_6*w_14) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Expression Object: w_33 := (x_7*x_15) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Expression Object: w_35 := (x_7*w_16) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Expression Object: w_37 := (w_8*x_17) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Expression Object: w_39 := (w_8*w_18) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 Initializing auxiliaries
 Problem size after  reformulation: 36 variables (0 integer), 10
 constraints.
 objectives:
 min w_20
 constraints:
 w_21 = 0
 w_23 = 0
 w_26 = 0
 w_28 = 0
 w_29 = 0
 w_32 = 0
 w_33 = 0
 w_36 = 0
 w_37 = 0
 w_40 = 0
 variables:
 x_0 [ -1e+50 , 1e+50 ]
 x_1 [ 0 , 1e+50 ]
 x_2 [ 0 , 1e+50 ]
 x_3 [ -1e+50 , 2 ]
 x_4 [ -1e+50 , 2 ]
 x_5 [ -1e+50 , 1e+50 ]
 x_6 [ 0 , 2 ]
 x_7 [ 0 , 2 ]
 w_8 (r:2, m:1) := (16+x_0-x_1-x_2-x_3-x_4-10*x_5-x_6-x_7) [ 0 , 2 ]
 w_9 (r:2, m:1) := (1-x_0+3*x_1+x_2) [ 0 , inf ]
 w_10 (r:2, m:1) := (-2-x_0+x_1+3*x_2) [ 0 , inf ]
 w_11 (r:2, m:1) := (4+x_0-3*x_3-x_4) [ 0 , inf ]
 w_12 (r:2, m:1) := (10+x_0-x_3-3*x_4) [ 0 , inf ]
 x_13 [ 0 , 1e+50 ]
 w_14 (r:3, m:1) := (x_0-3*x_6-x_7-w_8+x_13) [ 0 , inf ]
 x_15 [ 0 , 1e+50 ]
 w_16 (r:3, m:1) := (4+x_0-x_6-3*x_7-w_8+x_15) [ 0 , inf ]
 x_17 [ 0 , 1e+50 ]
 w_18 (r:3, m:1) := (10+x_0-x_6-x_7-3*w_8+x_17) [ 0 , inf ]
 w_19 (r:2, m:1) := (x_0^2) [ 0 , inf ]
 w_20 (r:3, m:1) := (1-2*x_0+w_19) [ -inf , inf ]
 w_21 (r:3, m:1) := (x_1*w_9) [ 0 , 0 ] integer
 w_23 (r:3, m:1) := (x_2*w_10) [ 0 , 0 ] integer
 w_25 (r:3, m:1) := (x_3*w_11) [ -inf , inf ]
 w_26 (r:4, m:1) := (2*w_11-w_25) [ 0 , 0 ] integer
 w_27 (r:3, m:1) := (x_4*w_12) [ -inf , inf ]
 w_28 (r:4, m:1) := (2*w_12-w_27) [ 0 , 0 ] integer
 w_29 (r:2, m:1) := (x_6*x_13) [ 0 , 0 ] integer
 w_31 (r:4, m:1) := (x_6*w_14) [ 0 , inf ]
 w_32 (r:5, m:1) := (2*w_14-w_31) [ 0 , 0 ] integer
 w_33 (r:2, m:1) := (x_7*x_15) [ 0 , 0 ] integer
 w_35 (r:4, m:1) := (x_7*w_16) [ 0 , inf ]
 w_36 (r:5, m:1) := (2*w_16-w_35) [ 0 , 0 ] integer
 w_37 (r:3, m:1) := (w_8*x_17) [ 0 , 0 ] integer
 w_39 (r:4, m:1) := (w_8*w_18) [ 0 , inf ]
 w_40 (r:5, m:1) := (2*w_18-w_39) [ 0 , 0 ] integer
 end

 List of user-set options:

                                     Name   Value                used
              boundtightening_print_level = 4                     yes
                    branching_print_level = 4                     yes
                 convexifying_print_level = 4                     yes
                         delete_redundant = yes                   yes
                             max_cpu_time = 499.965               yes
                     nlp_failure_behavior = fathom                yes
                      problem_print_level = 6                     yes
                              use_semiaux = no                    yes
 NLP0012I
               Num      Status      Obj             It       time
 Location
 NLP0014I             1         OPT 1.640146e-235.5g        4 0.003999g
 Couenne: 42 cuts (29 row, 13 col) for linearization
 created Variable Object: x_0 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Variable Object: x_1 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Variable Object: x_2 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Variable Object: x_3 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Variable Object: x_4 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Variable Object: x_6 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Variable Object: x_7 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_8 := (16+x_0-x_1-x_2-x_3-x_4-10*x_5-x_6-x_7)
 with mid-point strategy [clamp=0.2, alpha=0.25]
 created Variable Object: w_8 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_9 := (1-x_0+3*x_1+x_2) with mid-point
 strategy [clamp=0.2, alpha=0.25]
 created Variable Object: w_9 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_10 := (-2-x_0+x_1+3*x_2) with mid-point
 strategy [clamp=0.2, alpha=0.25]
 created Variable Object: w_10 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_11 := (4+x_0-3*x_3-x_4) with mid-point
 strategy [clamp=0.2, alpha=0.25]
 created Variable Object: w_11 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_12 := (10+x_0-x_3-3*x_4) with mid-point
 strategy [clamp=0.2, alpha=0.25]
 created Variable Object: w_12 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Variable Object: x_13 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_14 := (x_0-3*x_6-x_7-w_8+x_13) with mid-point
 strategy [clamp=0.2, alpha=0.25]
 created Variable Object: w_14 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Variable Object: x_15 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_16 := (4+x_0-x_6-3*x_7-w_8+x_15) with mid-
 point strategy [clamp=0.2, alpha=0.25]
 created Variable Object: w_16 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Variable Object: x_17 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_18 := (10+x_0-x_6-x_7-3*w_8+x_17) with mid-
 point strategy [clamp=0.2, alpha=0.25]
 created Variable Object: w_18 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_21 := (x_1*w_9) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Variable Object: w_21 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_23 := (x_2*w_10) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Variable Object: w_23 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_26 := (2*w_11-w_25) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Variable Object: w_26 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_28 := (2*w_12-w_27) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Variable Object: w_28 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_29 := (x_6*x_13) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Variable Object: w_29 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_32 := (2*w_14-w_31) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Variable Object: w_32 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_33 := (x_7*x_15) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Variable Object: w_33 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_36 := (2*w_16-w_35) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Variable Object: w_36 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_37 := (w_8*x_17) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Variable Object: w_37 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 created Expression Object: w_40 := (2*w_18-w_39) with mid-point strategy
 [clamp=0.2, alpha=0.25]
 created Variable Object: w_40 with mid-point strategy [clamp=0.2,
 alpha=0.25]
 Couenne initialized (0.019 seconds).

 Coin0506I Presolve 6 (-23) rows, 3 (-38) columns and 13 (-75) elements
 Clp0006I 0  Obj 0.0001 Primal inf 0.0082958817 (1) Dual inf 1.9999999 (1)
 Clp0006I 0  Obj 0.0001 Primal inf 0.0082958817 (1) Dual inf 1e+10 (2)
 Clp0006I 3  Obj -0.015610312
 Clp0000I Optimal - objective value -0.015610312
 Clp0032I Optimal objective -0.01561031175 - 3 iterations time 0.002,
 Presolve 0.00
 Clp0000I Optimal - objective value -0.015610312
 NLP0014I             2         OPT 1.62459e-215.5g       23 0.027995g
 Clp0000I Optimal - objective value -0.015610312
   x_0.  x =          1, lb = -0.0156103, cutoff = 0-----------------
  [  0.968514,  0.982915] lb = -0.0156103 {fea=1,btr=0}
  [  0.968514,  0.973806] lb = -0.0110699 {fea=1,btr=0}
  [  0.968514,  0.969126] lb = -0.00197622 {fea=1,btr=0}
   x_0.  x =          1, lb = -0.0156103, cutoff = 0-----------------
  [   1.02257,   1.04035] lb = -0.0156103 {fea=1,btr=0}
  [   1.03474,   1.04035] lb = -0.0126518 {fea=1,btr=0}
  [   1.04102,   1.04035] lb = -8.06182e-05 {fea=0,btr=0}
   x_2.  x =          1, lb = -0.0156103, cutoff = 0-----------------
  [  0.990072,  0.994676] lb = -0.0156103 {fea=1,btr=0}
  [  0.990072,  0.991888] lb = -0.0112907 {fea=1,btr=0}
  [  0.990072,  0.990468] lb = -0.00299315 {fea=1,btr=0}
   x_2.  x =          1, lb = -0.0156103, cutoff = 0-----------------
     # x0 = 1 iter   0: [ +0.966578 ->  +0.983289 --    +1.0436] /\/\  [
 0.966578,  0.983289]<0.966578,1>=>     # x0 = 1 iter   1: [ +0.966578 ->
 +0.974934 --    +1.0436] /\/\  [  0.966578,
 0.974934]<0.966578,0.983289>=>     # x0 = 1 iter   2: [ +0.966578 ->
 +0.970756 --    +1.0436] /\/\  [  0.966578,
 0.970756]<0.966578,0.974934>=>     # x0 = 1 iter   0: [ +0.966578 --
 +1.0218 <-    +1.0436] /\/\  [    1.0218,    1.0436]<1,1.0436>=>     # x0
 = 1 iter   1: [ +0.966578 --    +1.0327 <-    +1.0436] /\/\  [    1.0327,
 1.0436]<1.0218,1.0436>=>     # x0 = 1 iter   2: [ +0.966578 --   +1.03815
 <-    +1.0436] /\/\  [   1.03815,    1.0436]<1.0327,1.0436>=>     # x2 = 1
 iter   0: [ +0.989505 ->  +0.994752 --   +1.01189] /\/\  [  0.989505,
 0.994752]<0.989505,1>=>     # x2 = 1 iter   1: [ +0.989505 ->  +0.992129
 --   +1.01189] /\/\  [  0.989505,  0.992129]<0.989505,0.994752>=>     # x2
 = 1 iter   2: [ +0.989505 ->  +0.990817 --   +1.01189] /\/\  [  0.989505,
 0.990817]<0.989505,0.992129>=>     # x2 = 1 iter    [   1.00606,
 1.01117] lb = -0.0156103 {fea=1,btr=0}
  [   1.00923,   1.01117] lb = -0.0125536 {fea=1,btr=0}
  [   1.01086,   1.01117] lb = -0.00282679 {fea=1,btr=0}
   x_3.  x =          1, lb = -0.0156103, cutoff = 0-----------------
  [-1.79769e+308,-1.34078e+154] lb = -0.0156103 {fea=0,btr=0}
  [-1.79769e+308,-1.15792e+77] lb = -0.0156103 {fea=0,btr=0}
  [-1.79769e+308,-3.40282e+38] lb = -0.0156103 {fea=0,btr=0}
     pruned by Probing
 Cbc0006I The LP relaxation is infeasible or too expensive
 Cbc0013I At root node, 0 cuts changed objective from -0.015610312 to
 -0.015610312 in 1 passes
 Cbc0014I Cut generator 0 (Couenne convexifier cuts) - 0 row cuts average
 0.0 elements, 2 column cuts (2 active)
 Cbc0001I Search completed - best objective 1e+50, took 0 iterations and 0
 nodes (0.08 seconds)
 Cbc0035I Maximum depth 0, 0 variables fixed on reduced cost
  Couenne convexifier cuts was tried 1 times and created 2 cuts of which 0
 were active after adding rounds of cuts

 Couenne finished. No feasible solution found.

 0: [ +0.989505 --   +1.00595 <-   +1.01189] /\/\  [   1.00595,
 1.01189]<1,1.01189>=>     # x2 = 1 iter   1: [ +0.989505 --   +1.00892 <-
 +1.01189] /\/\  [   1.00892,   1.01189]<1.00595,1.01189>=>     # x2 = 1
 iter   2: [ +0.989505 --    +1.0104 <-   +1.01189] /\/\  [    1.0104,
 1.01189]<1.00892,1.01189>=>     # x3 = 1 iter   0: [-1.79769e+308 ->
 -1.34078e+154 --   +1.01189] /\/\
 [-1.79769e+308,-1.34078e+154]<-1.79769e+308,1>=>     # x3 = 1 iter   1:
 [-1.79769e+308 -> -1.15792e+77 --   +1.01117] /\/\
 [-1.79769e+308,-1.15792e+77]<-1.34078e+154,1>=>     # x3 = 1 iter   2:
 [-1.79769e+308 -> -3.40282e+38 --   +1.01054] /\/\
 [-1.79769e+308,-3.40282e+38]<-1.15792e+77,1>=>
 }}}

-- 
Ticket URL: <https://projects.coin-or.org/Couenne/ticket/14#comment:2>
Couenne, a solver for non-convex MINLP problems <https://projects.coin-or.org/Couenne>
Couenne, a solver for non-convex MINLP problems



More information about the Couenne-tickets mailing list