How to solve an infeasible column error in a linear optimization problem?

Question

While trying to solve a linear optimization problem (say Scenario A) with CPLEX, I am getting Infeasible Column ys.IM...G....044 error as shown below: [![enter image description here][1]][1]

The objective function has less than sign.

The representation of the columns in the mathematical programming system (mps) file is given below:

COLUMNS

ys.IM...G....044  nRM.....G....044  +5.00000e+00
ys.IM...G....044  nTXDZ........044  -1.00000e+01
ys.IM...G....044  ccap              +1.56000e+01
ys.IM...G....044  cs.IM...G.aaa044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aab044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aac044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aad044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aae044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aaf044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aag044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aah044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aai044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aaj044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aak044  -1.00000e+01
ys.IM...G....044  cs.IM...G.aal044  -1.00000e+01
ys.IM...G....044  cs.IM...G.baa044  -1.00000e+01
ys.IM...G....044  cs.IM...G.bab044  -1.00000e+01
ys.IM...G....044  cs.IM...G.bac044  -1.00000e+01
ys.IM...G....044  cs.IM...G.bad044  -1.00000e+01
ys.IM...G....044  cs.IM...G.bae044  -1.00000e+01
ys.IM...G....044  cs.IM...G.baf044  -1.00000e+01
ys.IM...G....044  cs.IM...G.bag044  -1.00000e+01
ys.IM...G....044  cs.IM...G.bah044  -1.00000e+01
ys.IM...G....044  cs.IM...G.bai044  -1.00000e+01
ys.IM...G....044  cs.IM...G.baj044  -1.00000e+01
ys.IM...G....044  cs.IM...G.bak044  -1.00000e+01
ys.IM...G....044  cs.IM...G.bal044  -1.00000e+01
ys.IM...G....044  cs.IM...G.caa044  -1.00000e+01
ys.IM...G....044  cs.IM...G.cab044  -1.00000e+01
ys.IM...G....044  cs.IM...G.cac044  -1.00000e+01
ys.IM...G....044  cs.IM...G.cad044  -1.00000e+01
ys.IM...G....044  cs.IM...G.cae044  -1.00000e+01
ys.IM...G....044  cs.IM...G.caf044  -1.00000e+01
ys.IM...G....044  cs.IM...G.cag044  -1.00000e+01
ys.IM...G....044  cs.IM...G.cah044  -1.00000e+01
ys.IM...G....044  cs.IM...G.cai044  -1.00000e+01
ys.IM...G....044  cs.IM...G.caj044  -1.00000e+01
ys.IM...G....044  cs.IM...G.cak044  -1.00000e+01
ys.IM...G....044  cs.IM...G.cal044  -1.00000e+01
ys.IM...G....044  func....G.        +1.55218e+00
ys.IM...G....044  func              +1.55218e+00
ys.IM...G....044  funct...G....044  +1.19399e-01
ys.IM...G....044  funcT........044  +1.19399e-01

RHS bounds are as follows:

rhs1              s.IM....G.)..023  +0.00000e+00
rhs1              s.IM....G.)..024  +0.00000e+00
rhs1              s.IM....G.)..037  +0.00000e+00
rhs1              s.IM....G.)..044  +0.00000e+00

y refers to new installed capacity limits in a year. I refers to Total installed capacity of technology limits in a year. 044 refers to the year 2044. G is a country. I tried to increase the rhs1 bound for 2044 to a very large value. But still the model is infeasible.

This is strange, because the model is feasible for another scenario with same COLUMNS and RHS bounds settings. How can I detect the source of infeasibility in the mps file and resolve it? [1]: https://i.sstatic.net/bme7li4U.png

There may not be many people here who can read raw MPS. I certainly can't. My initial reaction is that there are a heck of a lot of lines ending -10.0. Can you not make a much simpler MRE with fewer columns that still demonstrates the error? — Martin Brown
– Martin Brown, Commented Jan 13 at 12:59
Your last screenshot needs to be converted to a text segment in your question. — Reinderien
– Reinderien, Commented Jan 13 at 13:02

TimChippingtonDerrick · Accepted Answer · 2025-01-14 23:36:03Z

0

You could try reading the MPS file and writing the model out as an LP format file which should be easier to read

answered Jan 14 at 23:36

TimChippingtonDerrick

2,0921 gold badge12 silver badges13 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

hbstha123 Jan 16 at 0:02

Could you suggest any way or packages to convert mps to lp, and how is it easier to detect infeasibility with lp file?

TimChippingtonDerrick Jan 17 at 9:00

Using the CPLEX standalone solver commandline, read the MPS file and then write the model as an LP file. So... 'read xxx.mps' then 'write xxx.lp' should work. It isn't any easier to detect infeasibility in any format or style of model file, but the LP format is easier for people to read and hence easier to understand and spot mistakes if there are any, or understand why the model is infeasible.

Collectives™ on Stack Overflow

How to solve an infeasible column error in a linear optimization problem?

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related