Hacker News Reader

Comment by nkh | Hacker News Reader

Why would it need to replace them? From the article, they claim they have found a way to reduce the upperbound faster when searching large Integer problems. I don't see how that effects the current searching process. All of these solvers you can enter in an upperbound yourself if you have knowledge of the problem and know a previous solution. So it seems if this is just a programmatic way of reducing the upper bound, it should fit right in with current approaches. What am I missing?

View on HN

It's a research paper. You can write a theoretical paper and let others apply it practically, which others can figure out the practical aspect and report results of benchmarks, or others can also build on the theory.

This paper only has 2 authors. The other solvers are probably applying technique specific tricks and speedups, and you're working with approximate optimization, it's not that easy to move everything over.

> This paper only has 2 authors.

So? I don't get the relevance of the author count.

It's quite easy to go tell other people what they should do with their time.

These researchers are in the business of improving algorithms. Implementing them in large industrial (or open source) code bases in a maintainable way -- and then actually maintaining that code -- is a different skillset, a different set of interestes, and as was pointed out, besides the point.

Either you believe their results, then be grateful. Someone (yoU!) can implement this. Or you don't. In which case, feel free to move on.

Your tone comes off as entitled.

> Implementing them in large industrial (or open source) code bases in a maintainable way -- and then actually maintaining that code -- is a different skillset, a different set of interestes,

You're making a very general point on how algorithm research and software development are two different things, which is of course true. However OP's question is genuine: a lot of research in OR is very practical, and researchers often hack solvers to demonstrate that whatever idea offers a benefit over existing solving techniques. There are no reason to believe that a good new idea like this one couldn't be demonstrated and incorporated into new solvers quickly (especially given the competition).

So the quoted sentence is indeed a bit mysterious. I think it just meant to avoid comment such as "if it's so good why isn't it used in cplex?".

>business of improving algorithms

You do realize that the solver companies are in exactly the same boat, right?

And given how much the licenses cost, I'd love a new player to show up and bring them down to a reasonable level.

Since version 8.0.3, SCIP is available under Apache 2.0 License:

> https://www.scipopt.org/index.php#news

So the new player to show up is here. :-)

no they're not. they're in the business of making their customers' problems solve fast and well. That's of course strongly related, but it is _not_ the same. An algorithm may well be (and this is what OP might be hinting at) be more elegant and efficient, but execute worse on actually existing hardware.

I don't think they're talking about a bound for the optimum objective value, but a theoretical upper bound for a covering radius related to a convex body and a lattice. The bound would be useful in a lattice-based algorithm for integer linear programming. I don't think there exists an implementation of a lattice algorithm that is practical for non-toy integer linear programming problems, let alone one that is competitive with commercial ILP solvers.

Every time an integer feasible point is found during the iterative process these algorithms use (branch and bound), you get a new upper bound on the global minimum. It’s not clear to me how these dynamically generated upper bounds highly specific to the particular problem relate to the upper bounds of a more general nature that R&R produce.

> upper bounds of a more general nature that R&R produce

If it's an upper bound, it should be pretty easy to plug into the existing stuff under the hood in these solvers. Can you provide my insight into how the R&R "Upper bound" is different and "more general in nature"?

They prove a new upper bound to a combinatorial quantity that controls the worst-case running time of an algorithm of Dadush, not an upper bound to the optimal value of a given ILP instance.

If they wanted to see their ideas work in practice, they could implement Dadush's algorithm in light of these new bounds, but this would be unlikely to outperform something like CPLEX or Gurobi with all their heuristics and engineering optimizations developed over decades.

Otherwise, and this is the sense of the quoted sentence, they could go deep into the bowels of CPLEX or Gurobi to see if their ideas could yield some new speed-up on top of all the existing tricks, but this is not something that makes sense for the authors to do, though maybe someone else should.

Honestly?

The search for the 'exactly optimal solution' is way overrated

I think you can get a moderately efficient solution using heuristics at 1/10 of the time or less

Not to mention developer time and trying to figure out which constraints make your problem infeasible. Especially as they get more complicated because you want to make everything linear

I agree, especially when considering that a model is also not reality.

However, what folks often do is find a Linear Solution quickly, then optimize on the Integer Solution, which gives you a gap that you can use to choose termination.

The vast majority of the United States power grid (many thousands of power plants) are optimized in auctions every hour for the next day and every 5 minutes on the operating day. Finding the globally optimal solution is pretty important for both fairness and not wasting billions of dollars each year. I'd agree with you for a lot of problems though, but keep in mind there are plenty where they need full optimality or within a tiny percentage from it.

Yes and your very complex linear branch and bound solution won't run in 5 min.

That's what I'm getting at

Reply on news.ycombinator.com