ModelCompare

The purpose of this project is to compare two optimization models in two different files.

Running

If you have 2 files like model1.lp and model2.lp in the folder test. You can compare them by running:

compare_models(
    file1 = "test/models/model1.lp",
    file2 = "test/models/model2.lp",
    outfile = "test/models/compare_lp.txt", 
    tol = 1e-3
)

Analyzing the results

The section VARIABLE NAMES says which variable belongs to each model.

	MODEL 1
		w
		x
		z_a
    MODEL 2
		p
		t
		z_10_
		z_6_
		z_7_
		z_8_
		z_9_
		z_a_1_
		z_a_2_

The variables w,x,z_a belongs only to the first model and p,t,z_10_,z_6_,z_7_,z_8_,z_9_,z_a_1_,z_a_2_ belongs only to the first model. If both models have the same variables, it doesn't show on the results.

In the objective section, it shows which coefficient are different.

	SAME VARIABLES
	y_1_
		 MODEL 1 => 3.0
		 MODEL 2 => 5.0
	y_2_
		 MODEL 1 => 3.0
		 MODEL 2 => 5.0
	d
		 MODEL 1 => 5.0
		 MODEL 2 => 1.0
	DIFFERENT VARIABLES:
	MODEL 1:
		w => 0.5
		x => 2.0
		z_a => 1.0
	MODEL 2:
		z_8_ => 1.0
		z_a_1_ => 1.0
		z_7_ => 1.0
		z_a_2_ => 1.0
		z_9_ => 1.0
		z_6_ => 1.0
		z_10_ => 1.0
		p => 3.0

The variable y_1_ has a coefficient 3.0 multiplying it in the model1 and 5.0 in the model 2.

In the bounds section, it shows which bounds are different.

DIFFERENT VARIABLES:
	MODEL 1:
		w => [30.0,Inf)
		x => [0.0,Inf)
		z_a => [25.0,25.0]

The variable w has bounds $[30.0,\infty)$. The z variable is fixed in 25.

In the constraint section it shows the coefficients of the variables that are different and if the constraint bounds are different.

You can only compare constraints with the same name, otherwise, it will be ignored.

CONSTRAINT: c1
	DIFFERENT VARIABLES:
	MODEL 1:
		z_4_ => 2.0
		z_2_ => 2.0
		z_5_ => 2.0
		z_3_ => 2.0
		z_1_ => 2.0
		z_a => 1.0
	MODEL 2:
		z_a_1_ => 2.0
		z_a_2_ => 1.0
	SETS
		MODEL 1: MathOptInterface.GreaterThan{Float64}(3.0)
		MODEL 2: MathOptInterface.EqualTo{Float64}(3.0)


CONSTRAINT: z_con
	DIFFERENT VARIABLES:
	MODEL 2:
		z_8_ => 1.0
		z_7_ => 1.0
		z_9_ => 1.0
		z_6_ => 1.0
		z_10_ => 1.0

On the constraint c1. In the first model, it has the variables z_1_,z_2_,z_3_,z_4_,z_5_,z_a_ on the constraint c1 and the second model doesn't have this variables in this constraint. Also, z_1_ has a coefficient of 2 on the constraint c1 and so on.

In the first model, the c1 constraint has a bound of Greater than 3 and on the second model Equal to 3.