Different formulations lead to different results

Liting Lu

unread,

Nov 10, 2014, 5:11:58 AM11/10/14

to yal...@googlegroups.com

Hello Johan and everyone,

I formulated the problem with 2 different forms, but the solver returned 2 different results. For example, when I wrote the Item in the cost function as following:

objective = objective +((c_p1*z{k} + c_p2 + c_n1*(u{k}(1)-z{k})*0.3108 - vp_inv{k}*u{k}(2))*d1 + d0{k})^2;

The programming ran without any error. But when I wrote the code as following:

objective = objective + mf{k}^2;

with constraint as:

constraints = [constraints, mf{k} >= (c_p1*z{k} + c_p2 + c_n1*(u{k}(1)-z{k})*0.3108 - vp_inv{k}*u{k}(2))*d1 + d0{k}];

or even as:

constraints = [constraints, mf{k} == (c_p1*z{k} + c_p2 + c_n1*(u{k}(1)-z{k})*0.3108 - vp_inv{k}*u{k}(2))*d1 + d0{k}];

the solver reported that the problem was infeasible.

Is it because， that the equality and inequality residuals made the problem tighter? Thx!

Liting

Johan Löfberg

unread,

Nov 10, 2014, 5:15:08 AM11/10/14

to yal...@googlegroups.com

Should work. Which solver are you using? What is displayed when you run it

Message has been deleted

Liting Lu

unread,

Nov 10, 2014, 5:54:27 AM11/10/14

to yal...@googlegroups.com

I use 'mosek' as my solver. When another constraint was met, YALMIP reported the problem is infeasible.

Before there is already this kind information:

Interior-point optimizer terminated. Time: 0.84.

Optimizer terminated. Time: 0.94

Interior-point solution summary

Problem status : UNKNOWN

Solution status : UNKNOWN

Primal. obj: 1.9934448525e+008 Viol. con: 1e+002 var: 0e+000

Dual. obj: 1.9934662024e+008 Viol. con: 2e-009 var: 5e-007

Optimizer summary

Optimizer - time: 0.94

Interior-point - iterations : 128 time: 0.84

Basis identification - time: 0.00

Primal - iterations : 0 time: 0.00

Dual - iterations : 0 time: 0.00

Clean primal - iterations : 0 time: 0.00

Clean dual - iterations : 0 time: 0.00

Clean primal-dual - iterations : 0 time: 0.00

Simplex - time: 0.00

Primal simplex - iterations : 0 time: 0.00

Dual simplex - iterations : 0 time: 0.00

Primal-dual simplex - iterations : 0 time: 0.00

Mixed integer - relaxations: 0 time: 0.00

Mosek error: MSK_RES_TRM_STALL ()

At the infeasible point, the display is:

Interior-point optimizer terminated. Time: 0.53.

MOSEK PRIMAL INFEASIBILITY REPORT.

Problem status: The problem is primal infeasible

Optimizer terminated. Time: 0.62

Interior-point solution summary

Problem status : PRIMAL_INFEASIBLE

Solution status : PRIMAL_INFEASIBLE_CER

Dual. obj: -2.1407929344e+008 Viol. con: 8e-005 var: 4e-002

Optimizer summary

Optimizer - time: 0.62

Interior-point - iterations : 73 time: 0.53

Basis identification - time: 0.00

Primal - iterations : 0 time: 0.00

Dual - iterations : 0 time: 0.00

Clean primal - iterations : 0 time: 0.00

Clean dual - iterations : 0 time: 0.00

Clean primal-dual - iterations : 0 time: 0.00

Simplex - time: 0.00

Primal simplex - iterations : 0 time: 0.00

Dual simplex - iterations : 0 time: 0.00

Primal-dual simplex - iterations : 0 time: 0.00

Mixed integer - relaxations: 0 time: 0.00

Vollständigen Inhalt anzeigen

Johan Löfberg

unread,

Nov 10, 2014, 5:57:30 AM11/10/14

to yal...@googlegroups.com

You would have to supply reproducible code. Looks badly scaled to me

Liting Lu

unread,

Dec 9, 2014, 7:47:55 AM12/9/14

to yal...@googlegroups.com

Johan Löfberg

unread,

Dec 9, 2014, 9:23:54 AM12/9/14

to yal...@googlegroups.com

run yalmiptest(sdpsettings('debug',1)) and the error will be explicit

Or simply solve an LP and see how it crashes

sdpvar x
optimize(x>=0,x,sdpsettings('debug',1))

Reply all

Reply to author

Forward