linprog - infinite bounds
Luca
Details: consider a given linear program where some variables are unbounded, either below (lb(i)=-Inf for some i) and/or above (ub(i)=Inf for some i). Sometimes an "unbounded problem" message occurs if i set relevant bounds to Inf/-Inf, as above. if i try with some "big" number instead, the problem is solved correctly, and with no active big bounds suggesting a possible unboundedness.
I noticed that in the old, command-line help the usage of Inf/-Inf was suggested:
>> help linprog
LINPROG Linear programming.
X = LINPROG(f,A,b) attempts to solve the linear programming problem:
min f'*x subject to: A*x <= b
x
X = LINPROG(f,A,b,Aeq,beq) solves the problem above while additionally
satisfying the equality constraints Aeq*x = beq.
X = LINPROG(f,A,b,Aeq,beq,LB,UB) defines a set of lower and upper
bounds on the design variables, X, so that the solution is in
the range LB <= X <= UB. Use empty matrices for LB and UB
if no bounds exist. Set LB(i) = -Inf if X(i) is unbounded below;
set UB(i) = Inf if X(i) is unbounded above.
etc....
whereas in the new, "visual" help this seggestion has disappeared?
Is it because it may prove faulty?
Any suggestion?
B.t.w., my test problem is very small; yet it is causing me such strange problems.
Thanks in advance and anyway,
luca
Matt J
>
>
> whereas in the new, "visual" help this seggestion has disappeared?
"visual help"? you mean the html page "doc linprog"?
> Is it because it may prove faulty?
> Any suggestion?
==================
I'm not sure what's happening, but as a test, you could try to re-express the bounds using Aeq and beq instead of ub and lb and see if that makes a difference.
>
> B.t.w., my test problem is very small; yet it is causing me such strange problems.
=========
If it's small, perhaps you could show it to us...
Matt J
I meant, re-express the bounds using A and b rather than with Aeq and beq.
Luca
Thx Matt J for the suggestion.
This was something planned but delayed, since costed a bit of recoding. Nonetheless, i did it, leaving out those bounds with Inf/-Inf, since they are useless. So only the "meaningful" bounds have trurned to inequality constraints. Now i get an unfeasible problem. At this point, i realize that Inf/-Inf is not the issue, but probably the overall setting of my problem needs being revised. The physical problem from which everything originates is "feasible" (i.e., existence of solution) & "bounded" (i.e., finiteness of solution). I have to investigate in other directions.
So, i suppose this thread is to be declared closed
thx anyway for the help.
luca