Stack Exchange
log in sign up chat discuss faq users badges

Area 51 » Science

definition commitment beta
212
followers

This proposal is now in the Beta phase — example questions are locked!

65 Example Questions (2 closed)

active newest votes
up vote 21 down vote
Is the TSP polynomially solvable on certain classes of graphs?
added by LarrySnyder610 Mar 31 at 17:32
link
up vote 14 down vote
How can I compare the strength of different mixed-integer linear programming formulations?
added by Thiago Serra Mar 31 at 18:47
link
up vote 11 down vote
Under what conditions is the Lagrangian duality gap for an IP guaranteed to equal 0?
added by LarrySnyder610 Mar 31 at 17:30
link
up vote 11 down vote
What are the advantages/disadvantages of "generic" callbacks over "legacy" callbacks in CPLEX?
added by prubin Mar 31 at 18:14
link
up vote 11 down vote
Under what conditions can matrix rank constraints be modelled using mixed-integer convex programming?
added by rcorywright Mar 31 at 22:54
link
up vote 11 down vote
Can one use Reinforcement Learning for inventory optimization problem of multi echelon networks?
added by Afshin Oroojlooy Apr 1 at 5:41
link
up vote 10 down vote
Why are inventory optimization models usually solved with special-purpose algorithms instead of general-purpose nonlinear solvers?
added by LarrySnyder610 Mar 31 at 17:31
link
up vote 10 down vote
What are the practical/algorithmic implications of Farkas's lemma?
added by LarrySnyder610 Mar 31 at 17:32
link
up vote 10 down vote
What does it mean for a Markov chain to be ergodic?
added by LarrySnyder610 Mar 31 at 17:34
link
up vote 10 down vote
Does anyone have any examples of binary programming problems where binary decision diagrams provide tighter bounds than LP relaxations do?
added by prubin Mar 31 at 18:24
link
up vote 10 down vote
What are examples of optimization solvers/modeling languages for exponential cones?
added by Christian Kroer Mar 31 at 18:44
link
3  
Downvoting this question because guidelines say we should "Avoid asking questions that are subjective, argumentative, or require extended discussion. Stack Exchange does not work well with questions like 'Which is the best…'". area51.stackexchange.com/faq Feel free to reword the question. – LarrySnyder610 Apr 1 at 19:19
1  
@LarrySnyder610 Has this question been edited subsequent to your comment? It seems like a perfectly fine question as currently shown, and I could give a great and comprehensive answer to it, to include adding material for "matrix exponential" cone (using CVXQUAD). Even if it asked for which was best, that doesn't seem like the crime of the century to me, so perhaps I'm not Operations Research and Analytics Stack Exchange material. – Mark L. Stone Apr 2 at 12:16
@Mark L. Stone yes, the question-asker revised the question after my comment. The original question was something like "What is the best solver for ..." So the question is good as it is now, in my opinion. About asking "what's best," I agree it's not the crime of the century, but SE discourages it, and I think the more we can stick to SE guidelines and norms, the better chance we have. But I disagree you are not OR&A SE material -- I think SE works best when there are multiple voices and multiple opinions/answers. Please stay with us! – LarrySnyder610 Apr 2 at 14:33
1  
@LarrySnyder610 There is such a thing on SE as Community Wiki. Many of those questions and answers are very useful. "What's the best" questions can fall within Community Wiki in my opinion. – Mark L. Stone Apr 2 at 14:52
2  
Yes, as @LarrySnyder610 says, I changed the question to not say "best." I personally think that "best" was maybe even better; if I am going to an OR Stack Exchange I am often looking for opinions from people with more experience dealing with, e.g., exponential cones in practice. Still, making the question more generic by just asking for options seems fine. – Christian Kroer Apr 2 at 20:09
show 2 more improvement suggestions
up vote 10 down vote
Given a set covering problem with high density (say, 60% of coefficients in a typical constraint are 1), are there any useful bound-tightening cuts?
added by prubin Mar 31 at 18:47
link
up vote 10 down vote
In MCDM, what is a good way to convey a sense of the Pareto frontier to a decision maker when the number of criteria is "large" (say 10+)?
added by prubin Mar 31 at 19:32
link
up vote 10 down vote
Can two-stage stochastic programming be applied to short-term multi-period problems when the uncertain parameters are non-stationary?
added by Ehsan Mar 31 at 20:33
link
up vote 10 down vote
What are the conditions that the base-stock policy is optimal on a serial multi-echelon supply network?
added by Afshin Oroojlooy Mar 31 at 23:42
link
up vote 10 down vote
Does operations research play a role in the development of machine learning？
added by Ying Rong Apr 1 at 1:30
link
1  
This question would be stronger if it were more specific, e.g., "what are the ways that nonconvex optimization is used in ML" or something like that. – LarrySnyder610 Apr 1 at 19:21
up vote 10 down vote
How Reinforcement learning and Markov decision processes are related?
added by S. Phil Kim Apr 1 at 2:09
link
up vote 10 down vote
How do I use COIN-OR (or another open source integer programming solver) to model and solve a matching problem in C++?
added by akazachk Apr 1 at 2:18
link
1  
Downvoting this question because guidelines say we should "Avoid asking questions that are subjective, argumentative, or require extended discussion. Stack Exchange does not work well with questions like 'Which is the best…'". area51.stackexchange.com/faq Feel free to reword the question. – LarrySnyder610 Apr 1 at 19:19
1  
That makes sense. I have edited the question to make it concrete, something a student / researcher / practitioner may ask. – akazachk Apr 1 at 20:52
1  
For posterity, my old question was "What are the open source integer programming solver alternatives with the best performance?" This could be a good Community Wiki question, after reading the discussion on Christian's post. Though thinking a bit more on it, I think the original question was not so subjective / argumentative. It may require some extended discussion, but there are commonly accepted benchmarks that can be referred to. It would still be good to make it a Community Wiki in order to make it easier to update over time, since a normal "question" will have a fixed accepted answer. – akazachk Apr 2 at 15:57
1  
I think the original question is good. Make it Community Wiki or whatever, but I think this kind of question should be encouraged, not discouraged. Let people answer with thieir opinions and preferences, with whatever explanation they care to provide. Operations Research itself doesn't lend itself to clean unambiguous answers. – Mark L. Stone Apr 3 at 12:02
up vote 10 down vote
Is there a way to generate hard instances of the symmetric traveling salesman problem?
added by akazachk Apr 1 at 2:35
link
up vote 10 down vote
How machine learning and reinforcement learning can be used in supply chain problems?
added by Afshin Oroojlooy Apr 1 at 5:38
link
This question would be stronger if it were more specific, e.g., "are there examples of RL being used in real-world supply chains" or something like that. – LarrySnyder610 Apr 1 at 19:21
up vote 10 down vote
What could be a possible application of Reinforcement Learning for power grid problems?
added by Afshin Oroojlooy Apr 1 at 5:42
link
up vote 10 down vote
Where can I find a real dataset with retail demand and order quantities, including details like product hierarchy and order vs open-to-buy dates?
added by user201324 Apr 1 at 6:14
link
up vote 10 down vote
How should I choose between Generic Constraints and Big-M formulations in Gurobi?
added by Simon, edited by warren Apr 2 at 14:17
link
up vote 10 down vote
Are there examples where SAT solvers are competitive for optimization problems?
added by Simon Apr 1 at 7:11
link
up vote 10 down vote
How can I formulate this Packing Problem as ILP?
added by Simon Apr 1 at 7:15
link
up vote 10 down vote
In cplex, Benders decomposition is now in-built. Can I add surrogate constraints to avoid feasibility cuts like we do in regular BD implementation?
added by Jyotirmoy Apr 1 at 11:24
link
up vote 10 down vote
How do you select the most suitable uncertainty set for a Robust Optimization Formulation?
added by Vedat BAYRAM Apr 1 at 12:25
link
up vote 10 down vote
How do you obtain the dual of a second order cone programming model and how do you get the duals for such a problem in Java/Cplex?
added by Vedat BAYRAM Apr 1 at 12:28
link
up vote 10 down vote
Is EOQ inventory model used in any company in the real-world?
added by Afshin Oroojlooy Apr 1 at 13:16
link
Or better "How EOQ inventory model should be modified to be used realistically in any company?" – abhiieor Apr 3 at 10:46
up vote 10 down vote
What characteristics are important when choosing a meta-heuristic for a vehicle routing problem?
added by E. Tucker Apr 1 at 13:38
link
2  
This question would be stronger if it were more specific, e.g., "what characteristics are important when choosing a meta-heuristic for a vehicle routing problem" or something like that. – LarrySnyder610 Apr 1 at 19:23
1  
Edited. Thanks. – E. Tucker Apr 2 at 13:11
up vote 10 down vote
What is the difference between user cuts and lazy constraints?
added by rocarvaj Apr 1 at 14:37
link
up vote 10 down vote
How can I use GCG (part of SCIP) together with Gurobi?
added by Simon Apr 1 at 15:15
link
1  
Please use full name of CGC. I know and used Gurobi but can't remember what is CGC? – abhiieor Apr 3 at 11:04
up vote 10 down vote
What is the difference between constraint programming and integer programming?
added by Thiago Serra Apr 1 at 17:33
link
up vote 10 down vote
Are there any guidelines to improve the performance of difficult mixed integer programs?
added by EhsanK Apr 1 at 18:59
link
up vote 10 down vote
When is computing a tighter M for a big-M constraint worth the extra effort?
added by akazachk Apr 1 at 22:21
link
up vote 10 down vote
When using CPLEX, are there any differences in performance between APIs? Any other major differences?
added by CamiloOA Apr 2 at 1:00
link
up vote 10 down vote
How do you model "x1=1 or x2=1" using mixed-integer programming?
added by Michael Apr 2 at 13:52
link
Do you mean "How do I model a single exclusive-OR choice or a single inclusive-OR choice in mixed-integer programming?" If this is not what you mean, could you write a more clear question? – alerera Apr 3 at 13:39
up vote 10 down vote
How can I handle and exclude symmetries in my MIP?
added by YukiJ Apr 2 at 15:04
link
up vote 10 down vote
Are there papers or tutorials on the use of stochastic or robust optimization to train classification or prediction models?
added by Nathan Brixius Apr 2 at 16:19
link
up vote 10 down vote
Are there examples of medium or large-sized VRP that are solved optimally in real-world supply chains?
added by EhsanK Apr 2 at 16:20
link
1 2