Originally posted by: NguyenHuuTRi

Hi,

I do not know why anytime I use the below constraint my model can not return the result. 

forall (t in timeslot: t<7, g in group){
 (z[g][t]!= z[g][t+6]) => y[g][t]== 2;

 (z[g][t] == z[g][t+6])&&(z[g][t]==0) => y[g][t] == 1;

 (z[g][t] == z[g][t+6])&&(z[g][t]==1) => y[g][t] == 0;
};

I am getting stucked with my model. Could you please help me resolve this issue.

Thanks.,

 

 

 

range type = 1..14;
range course = 1..518; 
range room = 1..150;
range lecture = 1..353;
range timeslot = 1..12;
range group=1..252;
int lecturetime[lecture][timeslot] = ...;
int courselecture[course][lecture]=...;
int caparoom[type]= ...;
int coursegroup[course][group]=...;

 

tuple lecture_time { 
    int lecture; 
    int time; 
}; 

setof (lecture_time) LectureTimes = 
        {<l,t> | l in lecture,t in timeslot : lecturetime[l][t]==1};

 

tuple course_lecture { 
    int course; 
    int lecture;     
}; 
setof (course_lecture) CourseLectures = 
        {<c,l> | c in course, l in lecture : courselecture[c][l]==1};

tuple course_type{
    int idcourse;
    int idtype;
};
setof (course_type) coursetypes = ...;

tuple course_group { 
    int idcourse; 
    int idgroup; 
};

setof (course_group) CourseGroups = {<c,g> | c in course, g in group: coursegroup[c][g]==1};

tuple lecture_course_time {
    int lecture;
    int course;
    int type;
    int time;

    

}
setof (lecture_course_time) LectureCourseTimes = 
        {<l,c,ty,t> | <c,l> in CourseLectures, <l,t> in LectureTimes , <c,ty> in coursetypes};
    

dvar boolean x[LectureCourseTimes];
dvar boolean z[group][timeslot];
dvar boolean y[group][timeslot];

//dexpr int o1=sum (w in LectureCourseTimes) x[w];
//dexpr int o2=sum(t in timeslot:t<7, g in group) y[g][t];

//maximize staticLex(o1,o2);

maximize sum (w in LectureCourseTimes) x[w];

//minimize sum(t in timeslot:t<7, g in group)  y[g][t];

subject to
{ 

forall (<l,t> in LectureTimes) {
        assignLecture :    sum(<l,c,ty,t> in LectureCourseTimes) x[<l,c,ty,t>] <=1;

    };        

forall (c in course) {
        asignCourse: sum(<l,c,ty,t> in LectureCourseTimes) x[<l,c,ty,t>]<= 1 ; 
    };    
    

forall (t in timeslot, ty in type) {
        asignCapaRoomType: sum(<l,c,ty,t> in LectureCourseTimes) x[<l,c,ty,t>]<= caparoom[ty]; 
    };

 

forall(t in timeslot, g in group){
sum(<l,c,ty,t> in LectureCourseTimes)x[<l,c,ty,t>]*coursegroup[c][g]==z[g][t];
};

/*
forall (t in timeslot: t<7, g in group){
 (z[g][t]!= z[g][t+6]) => y[g][t]== 2;

 (z[g][t] == z[g][t+6])&&(z[g][t]==0) => y[g][t] == 1;

 (z[g][t] == z[g][t+6])&&(z[g][t]==1) => y[g][t] == 0;
};
*/
forall(t in timeslot, g in group){
sum(<l,c,ty,t> in LectureCourseTimes)x[<l,c,ty,t>]*coursegroup[c][g]<=1;
};

};

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Stanko_86

As variable y only appears in this constraint and has no influence on rest of your output variables, I suggest going with workaround where y will be calculated in post-process execution 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: NguyenHuuTRi

Hi,

 

Could you explain in details what I should do as I am very confused about what to do now.

 

Thanks.

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Stanko_86

execute
{
    for(var t in timeslot)
    {
        for (var g in group)
        {
            var yy = 0;
            if(t < 7)
            {
                if(z[g][t]!= z[g][t+6])
                {
                    yy = 2;
                }
                else if(z[g][t] == 0)
                { 
                    yy = 1;            
                }
                else if(z[g][t] == 1)
                {
                   yy = 0;                                
                }  
                writeln(t + "," + g + "," + yy);            
            }
        } 
    }     
}

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: NguyenHuuTRi

Hi,

As I have to minimize y[g][t] in my model, so If I use the syntax suggested , will it be possible?

 

Thanks.

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Stanko_86

As I mentioned in previous comment, I have noticed that y variable has not been used in other constrains or objective function except you mentioned, and based on that, I suggested the workaround how to calculate y.

In case that you want to use y in objective function, than my suggestion is not usefull. 

Also, I have noticed that you defined  y as a boolean, but you have something like y == 2?

In that case, y should be defined as dvar int y in 0..2

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: NguyenHuuTRi

Hi Sanko,

 

Thanks for your help.

Anyway, I wonder where the value of y could be extracted from the model as I have run the model but there is no value of y at all.

 

Thanks.

#DecisionOptimization
#OPLusingCPLEXOptimizer