Jaromił Najman

Hi Stefan,You are allowing for an approximation error of at most 20 by setting FuncPieceError = 20. This parameter should be set to something like FuncPieceError=1e-3 or a bit lower. Note that Guro...

None of the models reaches an ordering time of ~2500s as seen above. Why does it often take so much longer to order when using the automatic value for BarOrder (= -1) than selecting BarOrder = 0 o...

m.addConstr(sum(flow_bin[i,j,t] for i in inland_terminal if i!=j) >= 0.5 + (M*b) - b) The right hand side should read \(\texttt{0.5 + (M*b) - M}\). Another question is, why the 0.5? Is that beca...

The only difference is than \(L\) is \(N\) and \(f_i \in \mathbb{N}\) in your formulation, is this correct? Then no, nothing changes compared to my last response.

Terminal5 receives a total of 12 containers where 9 are from port and 3 are from terminal8(which should not happen), whereas it should receive only 6(either from one terminal or port). Then it tak...

Currently each solve+creation takes roughly 0.004-0.005 seconds, and I am wondering if that is a kind of limit (Python API, would using the C++ API be noticeably superior here?) Which of the two ...

Hi Hussein, What does it mean when gurobi uses the Presolve, and now the Presolve removes all binary variables? Does that mean the problem can be formulated without binary variables?  Theoretical...

Question 1: When modelling the binary constraint for flow, should I model it as this  for t in time:     m.addConstr((flow_bin.sum(i,j,t) <= 1)) or as this for t in time:   m.addConstr((flow_bin...

In this part, the I want to constrain the flow from inland terminals. Such that only one of them should be able to provide containers. If I were to do it for every i j in nodes then port will be i...

However, will I have to do it individually if the problem has many more variables? You somehow have to provide this information to Gurobi. If you have the optimal solution point as some data stru...