Why does the [x,fval,ex​itflag,out​put,lambda​] = linprog(f,​A,b,Aeq,be​q,lb,ub) function return only integer numbers?

Hi, I'm doing an optimization problem where I use the [x,fval,exitflag,output,lambda] = linprog(f,A,b,Aeq,beq,lb,ub) funtion. However, the results I get give me only integer numbers. For example, I'm expecting a result like 1.710, but I get 2.0. When I expect a value of 0.855, I get 1.0

How can I get the rational numbers with decimals, instead of only integer numbers?

[Torsten]

Does your "expected result" give a smaller value for f than the "integer result" linprog returns ?

[Romeo Tahal]

My f was defined as [10 12]'. The result of the linprog returns the 10 perfectly, which is the LMP at bus 1. For LMP at bus 2 I should get 12, but instead the linprog returns a value higher than 12.

[Steven Lord]

Does your non-integer solution satisfy the constraints represented by the A, b, Aeq, and beq matrices?

Does it satisfy the bounds represented by the lb and ub vectors?

If the answer to either of those questions is no, then your non-integer "solution" is not a solution.

[Torsten]

@Romeo Tahal

My f was defined as [10 12]'. The result of the linprog returns the 10 perfectly, which is the LMP at bus 1. For LMP at bus 2 I should get 12, but instead the linprog returns a value higher than 12.

You should not get f as solution from linprog, but an x that minimizes f'*x under your constraints.

So again the question: is [10 12]*result_linprog or [10 12]*(your result) smaller ?

Does your result satisfy the constraints ?

Does it satisfy the bounds ?

[Romeo Tahal]

Yes, my x satisfies the constraints, and also the bounds.

[Torsten]

And what about the third question:

is [10 12]*result_linprog or [10 12]*(your result) smaller ?

[Romeo Tahal]

The results are slightly higher than expected, but I can live with that.

[Torsten]

If you know a solution you can live with, you don't need an optimizer, do you ?

Answer 1

[Sarthak]

Hi Romeo,

I understand that you want to set the precision of the output variable.

You can use the format function to set the precision of the output.

format long  
x,fval,exitflag,output,lambda = linprog(f,A,b,Aeq,beq,lb,ub);

To know more information about the “format” function, you can visit the documentation page of format.

I hope this solves your query.