Not signed in (Sign In)

Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.

    • CommentAuthorgalileopro
    • CommentTimeApr 17th 2012
    Hello, MTS.
    My topics related to the optimization of the stop-loss reinsurance cover, due to mismatch of mathematical topics. But this question relates to actuarial mathematics. I did learn in the course of actuarial mathematics. Why is the administration closes all my threads? I am very good reviews on this forum, I advised him a good specialist in the theory of probability and stochastic processes. I look forward to help. How do I formulate a question to match the theme forum? I have read the forum rules. There is a section of the theory of probability. Why is my question does not apply to mathematics?
    Maybe I can get help on a particular method of optimization?
    I really need to sort out this issue. I am a student. I can understand not all by himself. How can I restate the problem to obtain useful information?
    • CommentAuthorMariano
    • CommentTimeApr 17th 2012 edited

    The main problem with your questions is —I think— not the subject but their form.

    If you take the time to browse existing, non-closed questions you will immediately notice that they are not at all similar to your

    Interested in any information on optimizing the stop-loss reinsurance and optimization criteria. Interesting articles, links and literature.

    which reads like a telegraph message, and does not contain even one complete sentence.

    You should read the page carefully. If you have a question that involves real issues of research-level mathematics, you can ask it here, but you need to follow the advice on that page. However, as MTS suggested, may well be a more appropriate forum.


    It's a little hard to be sure about the topic because the questions are so vague and poorly written, but I think there is a pretty reasonable case that the OP is not asking about pure mathematics at a level appropriate to MO. A search on MathSciNet turns up plenty of articles with "stop-loss" somewhere in them, but essentially all are in journals that do not seem to be pure mathematics (MathSciNet, rightly, has a broad policy about what it indexes). Not to mention that "someone wrote an article about it in pure mathematics at some point" is necessary, but far from sufficient condition for being MO appropriate.


    Mariano, some people still pay or bandwidth usage! ;-)

    The questions are inappropriate first and foremost because they are not well-formulated mathematical question. I don't know if stop-loss methods can be considered research-level or not, but the point is that it is impossible to tell here because the question is not precise enough. If the OP had asked a question about a specific method, we could have easily decided whether that method was research-level or a well-known method. But this is not the case here.

    Moreover, MO is generally not a good place for reference requests. Only occasionally have I seen reference requests that were so specific that they generated interesting answers.
    @Ben - I don't think it has been decided that MO is only for pure mathematics.

    I certainly think research level questions in actuarial mathematics are appropriate here. However, I agree that general reference requests about a broad topic are not suitable.

    Just to clarify for the original poster, in case it is necessary - 'research level' in this context means 'level of people doing *original* research'. In other words, an appropriate question is one which is relevant to people working on questions to which no one knows the answer. These will usually be questions to which the answer, if it is known, is known only to a small community of specialists.

    I don't think it has been decided that MO is only for pure mathematics.

    That's fair; I was running off to teach and didn't write as carefully as I should have. That said, I think there does have to be some cutoff at which we say "I won't argue about whether it's mathematics, but it's not part of the bailiwick of MO." I think it's quite important that we stick to questions with research-level mathematical content. That doesn't mean that they have to be purely mathematical in nature, but if we're talking about a topic where many of the papers are in Insurance: Mathematics and Economics then it's not unreasonable to think we've gotten to the other side of that line.

    @Ben - I disagree with you about Insurance: Mathematics and Economics.

    We have never settled whether research in the end of applied mathematics not interested in proving theorems should be part of the remit of MO.

    A cursory look at Insurance: Mathematics and Economics shows it mostly publishes papers on original research in applied mathematics of that kind as well as some papers that prove theorems.

    > in the end of applied mathematics not interested in proving theorems >

    If someone does not care about proofs, I am not interested in addressing his or her questions. All the applied mathematicians I know are interested in proofs even if they have to come up with (usually tentative) answers that are not proved to be correct. If the "I don't care about having a proof" crowd takes over, I am out of here.

    @Bill - I know this is not exactly what you said, but I would be surprised if you have never seen this kind of work before.

    Several times I have seen colloquium or seminar talks (in a math department) which said roughly the following (all details made up by me).

    Delta Airlines wanted a solution to this 10,000 variable integer program that was guaranteed to be within 10% of optimal. Here is what the equations look like and here is how they relate to the problem of scheduling pilots for flights. The previous published record for integer programs of this kind was for 7000 variables. Using this clever combination of these various known techniques, we were able to get an answer in 8 months on this kind of computer.

    One could technically say that this person proved the theorem 'this particular solution to this particular integer program is within 10% of optimal' (since they are using algorithms whose deviation from optimal has proven bounds and some simple calculations show they combine to less than 10%), but that strikes me as a rather odd way of putting things, and I would not say this researcher is interested in proving theorems.

    Someone doing this kind of research might ask on MO: what is known about how far this particular kind of reduction can take you from optimal? Would such questions be welcome?

    I don't have a strong opinion; I am most interested in there being a clear sense of what the community thinks.
    • CommentAuthorgilkalai
    • CommentTimeApr 18th 2012
    One of the major revolutions in mathematics was the realization that many important applied problems do not have analytic answers, and that heuristic methods, numerics, simulation, and modeling are sometimes competing paradigms to the theorems/proof paradigm.
    So for sure there are important areas in applied mathematics where proving theorems is a secondary issue. Certainly, good research level questions in applied mathematics, even those which do not deal with theorem/proofs should be welcome on MO. (It will be inappropriate for the comminity of MO users to exclude mathematical areas which are not represented, explicitely or implicitely.)

    On the other hand taking into account how rigid are the MO questions in terms of mathematical areas overall, over these years, it is extremely unlikely that some problems outside the theorem/proofs paradign (or in other underrepresented area of math) will lead to the "I don't care about proofs" crowd taking over. How is this possible at all?
    What exactly is applied mathematics? Optimization and Game Theory seem to be accepted applied areas. What Alexander Woo just described seems to have a very different flavor.

    My view is consistent with Gil's even if the emphasis in my post is different.