Inequalities in MixedIntegerLinearProgram are not properly handled

[ppurka]

The <= and >= inequalities are not properly handled in MixedIntegerLinearProgram - essentially in both sage.numerical.mip.LinearConstraint and sage.numerical.mip.LinearFunction. Here is an example:

sage: p = MixedIntegerLinearProgram()
sage: b = p.new_variable()
sage: b[0] <= b[1] <= 2  # This is not ok
x_1 <= 2
sage: (b[0] <= b[1] <= 2).constraints # Not ok
[x_1, 2]
sage: b[0] >= b[1] # This is ok
x_1 <= x_0
sage: 1 >= b[1]    # This is not since it starts with a constant
False
sage: 1 >= b[1] >= 2*b[0] # Not ok
False
sage: b[2] >= b[1] >= 2*b[0]  # Not ok
2 x_0 <= x_1

CC: @dimpase

Component: linear programming

Keywords: MixedIntegerLinearProgram LinearConstraint LinearFunction

Reviewer: Dmitrii Pasechnik, Punarbasu Purkayastha

Issue created by migration from https://trac.sagemath.org/ticket/12546

[dimpase]

comment:1

this is a duplicate of #12091, right?

[ppurka]

comment:2

This is dupe of #12091. Please close. I don't see the option to close tickets. ( I got no powers :'( )

[dimpase]

comment:3

Replying to @ppurka:

This is dupe of #12091. Please close. I don't see the option to close tickets. ( I got no powers :'( )

Me neither, but we can positively review it to its quick and painless demise.

Hoogedelgeleerde Release Manager, AUB, dit is een dupe...

[nathanncohen]

comment:4

This + "positive review" is the standard incantation :-)

Nathann

[dimpase]

comment:5

Replying to @nathanncohen:

This + "positive review" is the standard incantation :-)

Nathann

natuurlijk, als u nu in Brussel woont, moet u ook de laatste woord, "dupe", ofwel "slachtoffer", kennen :)

[nathanncohen]

comment:6

natuurlijk, als u nu in Brussel woont, moet u ook de laatste woord, "dupe", ofwel "slachtoffer", kennen :)

My god. This last incantation is not in my book. Hopefully Jeroen will understand :-D

Nathann

[jdemeyer]

Reviewer: Dmitrii Pasechnik, Punarbasu Purkayastha