Originally posted by: Crubal

Hi,

 

I am defining the sets when building the mathematical model:

set1: a station (built) k is built on a path from nodes i to j; such that d(i, k) <= safety_distance /2 ;

set2: a station (built) k is built on a path from nodes i to j; such that d(k, j) <= safety_distance /2 ;

Where i, j are interchanges

 

I have the initial code:

// Define Stations and related parameter

 int Resources = ...; 
 range Stations = 1..Resources;

 range Interchanges = 1..Resources;

 

// Define Traffic Flow sets, distance and related binary variable

 tuple flow    {int i; int j;}
 setof(flow) Flows  = {<i,j> | ordered i,j in Interchanges};
 float distance_flow[Flows] = ...;

 float Safety_distance;

 

// Binary variable is defined this way whether a station is built or not 

 dvar int Build[Stations];

 

setof(flow) set1 = [<i,j> in Flows, k in Build[Stations] | distance_flow[<i,k>] <= 0.5 * Safety_distance];

In the last line, it has errors, and actually I don't know how to define this type of sets. 

 

 

 Besides, I am not sure how to define a binary variable: Cover[<i,j> in Flows]

 Where Cover[<i,j> in Flows] = 1 means in round trip (i, j), the flow is covered by a fueling station (Build[Stations] = 1); and 0 otherwise. 

 

 In my way, I just write : dvar int Cover[<i,j> in Flows]; 

 

Thank you for your help!

 

 

 

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

you can't define a set that will change according to a decision variable.

What you should do instead of writing s1= is to write something like

dvar boolean FlowsInSet1[Flows];

and then you will add constraints about what FlowsInSet1 means.

For Cover, you could declare

dvar boolean Cover[<i,j> in Flows] ;

and then add constraints like

subject to
 {
forall(<i,j> in Flows) Cover[<i,j>]==
(1<=Build[i]+Build[j]);
}

regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Crubal

Thanks Alex,

 

I am sorry for the late response, but it is useful! 

May I ask if I code like 

 {
forall(<i,j> in Flows) Cover[<i,j>]==
(1<=Build[i]+Build[j]);
}

 

How can I add safety_distance into the model? Since the constraint is valid only when the distance between i and j is lower than 0.5 * safety_distance.

 

Thank you!

 

Crubal

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: NALK

Dear ALEX,

 

I have a confusion about  how to define a set of arrays that is related to a decision variable , 

range q=1..N;

{string} Group={"S1","S2"}  

tuple element_Group {  int id;}

{element_Group} Group[q]=...;   ///  how to link data from excel or any other file to this array without using data file 

 

for all j in q,   I need to say  also  for all   m in set of j ( {}_j,  ),   can we do that using OPL;

Thanks a lot 

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

hi

how to link data from excel or any other file to this array without using data file 

==> you could use SheetRead, see the example oil

for all j in q,   I need to say  also  for all   m in set of j ( {}_j,  ),   can we do that using OPL;

==> yes

forall(j in q) forall(m in myset[j])

regards

 

https://www.linkedin.com/pulse/ibm-think-lia-loptimisation-et-moi-alex-fleischer/

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: NALK

Thank you ALEX , it is really helpful,

 

I have one more question  ,    forall( j in q)   forall(m in setof(j))   is that true that m moves in the setof(j) even its of different dimensions like that setof(j)=[ {1},{2,1000}];

for j=1,  m takes 1 as a value,

 

for j=2, m takes 2 and 1000 as values

 

is that true?

 

Thanks a lot again,

 

regarding the forum also, it will be deleted on Jan,2020 ,  is there any alternative plan for that,

 

Thanks

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi

yes you re right

and for the forum do not hesitate to comment or like https://www.ibm.com/developerworks/community/forums/html/topic?id=807057f2-57d6-42d6-a4b5-34dbce571c7b&ps=25

regards

#DecisionOptimization
#OPLusingCPLEXOptimizer