Re: A unbounded problem of a Quadratic conic optimization makes me so confused..

[Johan Löfberg]

Why would those constraints guarantee that the solution is bounded?

It's like saying [x + y <= 100]  guarantees that x and y are bounded

[Johan Löfberg]

Presumably, you are looking at a problem which has some kind of relevance. If so, you should be able to look at all variales and judge if the are relevant. If you encounter unboundness, you should be able to see on the solution that something is missing in terms of constraints.

If I optimize the development and sale of a car, and notice that the profit is unbounded, I should look at the solution (when I bound the profit to say 10000000000 to get a solution), and if I discover e.g. that the total wage cost is 0, I know where the problem is.