在CPLEX中,我构建了一个附件模型,该模型用于将产品定位到位置。试图按升序对产品分配到位置的约束Stackingorder [s]没有给我想要的结果。
如果将产品分配到位置xi,则其堆叠顺序+1应该比位置xi之后的所有产品小。
CPLEX似乎忽略了所有其他约束。
我应该如何更改约束或模型以使其起作用?
forall(w in Locations: w+1 in Locations, s in Products)
ctStackingorder:
{(Slot[s][w+1] * Stackingorder[s]) <= Slot[s][w] * (Stackingorder[s]+1);}
int Fixed = ...;
int NbLocations = ...;
range Locations = 0..NbLocations-1;
int NbProducts = ...;
range Products = 0..NbProducts-1;
int Capacity[Locations] = ...;
int LocationCosts[Products][Locations] = ...;
int RequiredLoc[Products] = ...;
int Stackingorder[Products] = ...;
dvar boolean Use[Locations];
dvar boolean Slot[Products][Locations];
dvar int SError[Products][Locations];
minimize
sum( w in Locations )
Fixed * Use[w] +
sum( w in Locations , s in Products )
LocationCosts[s][w] * Slot[s][w] +
sum( w in Locations , s in Products )
SError[s][w] *1000 * RequiredLoc[s];
subject to{
forall(s in Products )
ctProductHasEnoughLocations:
sum( w in Locations)
Slot[s][w] * Capacity[w] == RequiredLoc[s];
forall(s in Products, w in Locations: w+1 in Locations)
ctFacings:
if(RequiredLoc[s] >1){ Slot[s][w+1]==Slot[s][w];}
forall( w in Locations, s in Products )
ctUseSlotProduct:
Slot[s][w] <= Use[w];
forall( w in Locations )
ctMaxUseOfLocation:
sum( s in Products )
Slot[s][w] <= Capacity[w];
forall(w in Locations: w+1 in Locations, s in Products)
ctStackingorder:
{(Slot[s][w+1] * Stackingorder[s]) <= Slot[s][w] * (Stackingorder[s]+1);}
}
{int} Productsof[w in Locations] = { s | s in Products : Slot[s][w] == 1 };
execute
{
writeln("Open=",Use);
writeln("Storesof=",Productsof);
}
.dat
Fixed = 30;
NbLocations = 6;
NbProducts = 5;
RequiredLoc = [1,1,1,1,1];
Capacity = [1,1,1,1,1,1];
LocationCosts = [
[ 1, 1, 1, 1, 1, 1 ],
[ 1, 1, 1, 1, 1, 1 ],
[ 1, 1, 1, 1, 1, 1 ],
[ 1, 1, 1, 1, 1, 1 ],
[ 1, 1, 1, 1, 1, 1 ] ];
Stackingorder = [328,326,228,226,226];
答案 0 :(得分:0)
您的模型不可行,CPLEX放松了它。
如果您希望该约束成为硬约束,则应将其更改为
forall(s in Products )
//ctProductHasEnoughLocations:
sum( w in Locations)
Slot[s][w] * Capacity[w] == RequiredLoc[s];
(无标签)
然后其他约束将被放松。
请参见
最诚挚的问候