14
$\begingroup$

I would like to compare human solutions for OR-problems to optimal solutions. Therefore, I am looking for a dataset with the manual solutions computed by a human. I do not have a specific problem in mind, any problem would be nice.

Does such a dataset exist?

| improve this question | |
$\endgroup$
  • 4
    $\begingroup$ Probably not what you are looking for, but fun to mention: there has been a scientific experiment in which people were asked to maximize their winnings in the Monty Hall Dilemma. Interestingly, humans were outperformed by pidgeons. $\endgroup$ – Kevin Dalmeijer Aug 23 '19 at 15:03
9
$\begingroup$

By chance, I happened across this paper earlier today. It discusses human performance in solving the TSP manually. The paper is short but there are a few other references in it that probably provide more details of other experiments.

| improve this answer | |
$\endgroup$
9
$\begingroup$

From my experience with industry: typically, you do not obtain the human planners' solutions to compare against optimal solutions for at least two reasons:

  • the decisions/plans/solutions are not recorded, at least not completely
  • the manual plans sometimes violate constraints in your model.

Even if both can be dealt with, companies don't want their data, let alone plans available online, which is reason number three (and probably the show stopper).

| improve this answer | |
$\endgroup$
  • 1
    $\begingroup$ I believe that the reasons one and three you mention are indeed the main reasons why the data is not (widely) available. Nevertheless, the second reason actually is a good reason to compare human planners' solution against optimal solutions. It is interesting to see which constraints are violated by the human planners. It could that these violations are mistakes by the planners, but it could also very well be that your model does not represent the reality well enough. $\endgroup$ – Bgz6 Aug 28 '19 at 8:35
7
$\begingroup$

It is not a dataset of solutions, but Section 1.4 of the TSP book [1] has a discussion about "Human solution of the TSP" with several references to the relevant literature.

[1] Applegate, D.L., Bixby, R.E., Chvatal, V. and Cook, W.J., 2006. The traveling salesman problem: a computational study. Princeton university press.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.