So I made a model at work that I need to solve and my problem is it takes way too long to complete. I use NEOS servers since I can only run bonmin on my computer and I am thus limited to 8 hrs. I tried using BARON and COUENNE. This is my model:
set P;
param C;
param scie;
param l{P};
param pond{P}; #fraction de pouce
var x{1..C} integer;
var LIEN{P, 1..C} binary;
var START{P,1..C} binary;
var y{1..C} binary;
maximize Prod: sum {p in P, c in 1..C} (LIEN[p, c]*x[c]*pond[p])* y[c];
subject to SumX: sum {c in 1..C} (x[c]+scie)*y[c]<= 96-scie;
subject to SumLIEN{p in P} : sum{c in 1..C} (LIEN[p, c] * (x[c]+scie))*y[c]<=l[p];
subject to x1{c in 1..C}: x[1]=0;
subject to x_min{c in 2..C}: x[c]>=4;
subject to x_max{c in 1..C}: x[c]<=20;
subject to emptysaw{i in P, j in 1..C}: LIEN[i,1]=0;
subject to start1{p in P,c in 2..C}: START[p,c]>=LIEN[p,c]-LIEN[p,c-1];
subject to start2{p in P}: sum{c in 1..C} START[p,c] <=1;
This is a sample of 15 of my data. The real data is 120 long, but I could settle for 60:
param scie := 0.552;
param C:= 6;
set P :=
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
;
param l:=
1 3.948
2 13.448
3 17.448
4 21.448
5 25.448
6 29.448
7 33.448
8 37.448
9 41.448
10 45.448
11 49.448
12 53.448
13 57.448
14 61.448
15 66.448
;
param pond:=
1 0.0709
2 4.3669
3 13.5923
4 22.5757
5 21.1709
6 17.3977
7 10.9697
8 5.4999
9 2.6314
10 1.1629
11 0.3632
12 0.1159
13 0.0535
14 0.0159
15 0.0122
;
Any help would be greatly appreciated!