Decision Optimization

 View Only

Optimization for workforce scheduling with IBM ILog cplex 12.10

  • 1.  Optimization for workforce scheduling with IBM ILog cplex 12.10

    Posted Sun March 29, 2020 12:45 AM

    Originally posted by: Kylie gp


    Hello everyone. I am a new for using IBM ilog cplex and I have a lot of wondering about that.

    My objective output from cplex is to know that who worker works in which workstation with which machine in which period and day.

    That shows I have 5 indexes these are i = workers, j = workstaions, m = machines, k = periods and l = days. 

     

    I wonder that can I use 2 boolean decision variables which are X[i][j][k][l] and Y[m][j][k][l] or using a boolean decision variable with 5 indexes which is W[i][j][m][k][l] for solving this problems ?? I defined X will be 1 if worker (i) works in workstaion (j) in period (k) in day (l) and Y will be 1 if machine (m) works in workstaion (j) in period (k) in day (l).

     

    I have a  .mod model is (I am not sure it's correct but It can run the output)

    int Nbmachines =...; 
    int Nbworkers =...; 
    int Nbworkstations =...; 
    int Nbperiods =...; 
    int Nbdays =...;  
    int hazard =...; 

    range machines = 1..Nbmachines;
    range workers = 1..Nbworkers;
    range workstations = 1..Nbworkstations;
    range periods = 1..Nbperiods;
    range days = 1..Nbdays;

     

    float h [workstations] =...;
    float N [workstations] =...; Worker required per workstations
    float B [workers][workstations] =...;
    float A [machines][workstations] =...;
    float M [workstations] =...; //Machine required
    float P [machines] =...; // Power of machines
    float PM [workstations] =...; //Power required 


    //3 dimentions variable read from excel t[j][k][l]
    int t1[1..Nbworkstations*Nbperiods*Nbdays] =...;
    int tArray[j in 1..Nbworkstations, k in 1..Nbperiods,l in 1..Nbdays] = t1[l+Nbdays*(k-1)+Nbperiods*Nbdays*(j-1)];

    execute {
       writeln(tArray);
    };

     

    //Decision Variables
    dvar boolean X[workers][workstations][periods][days];
    dvar boolean Y[machines][workstations][periods][days];
    dvar boolean y[workers];
    dvar boolean z[machines];

     

    //Objective function
    dexpr float Objective = sum(i in workers) y[i];
    minimize Objective;


    //Contraints
    subject to { 
      //[OK] 1. limited hazard of the day for worker
            forall(i in workers) 
            sum(j in workstations,k in periods,l in days) h[j]*X[i][j][k][l] <= hazard
            
      //[OK] 2. Every workstations need to assign workers as worker required per workstations.
            forall(j in workstations, k in periods, l in days)
            sum(i in workers) X[i][j][k][l]== N[j]*t1[l+Nbdays*(k-1)+Nbperiods*Nbdays*(j-1)];
            
      //[OK] 3. A workstation needs One worker per one period and one day.  
            forall(i in workers, k in periods, l in days)
            sum(j in workstations) X[i][j][k][l] <= 1;
            
      //[OK] 4. Workers [i] are chosen from who be able to work in workstation j.
            forall(i in workers,j in workstations, k in periods, l in days)
            X[i][j][k][l] <= B[i][j];
            
      //[OK] 5. Workers [i] are chosen from all we have 30 workers.

            forall(i in workers,j in workstations, k in periods, l in days)
            X[i][j][k][l] <= y[i];
             
      //[HELP] 6.  machines [m] are chosen from all we have 56 machiens. 
            forall(m in machines, j in workstations, k in periods, l in days)
            Y[m][j][k][l] <= z[m];
      **I've got problem with 6. because the output does not show machines are chosen as I define 1 if they are chosen and 0 if other. I attached output of z[m] below.
            


    //[OK] 7. A workstation needs one machines per one period and one day.  
            forall(m in machines, k in periods, l in days)
            sum(j in workstations) Y[m][j][k][l] <= 1;
            

    //[OK] 8. Every workstations need to assign machines as machines required per workstations.
            forall(j in workstations, k in periods, l in days)
            sum(m in machines) Y[m][j][k][l]== M[j] * t1[l+Nbdays*(k-1)+Nbperiods*Nbdays*(j-1)];
     

    //[OK] 9. Every workstations need to assign machines that power of machines as power required per workstations.       
            forall(m in machines, k in periods, l in days)
            sum(j in workstations) Y[m][j][k][l] * PR[j] <= P[m];
            
    //[OK] 10. Machines [m] are chosen from which be able to work in workstation j.
            forall(m in machines, k in periods, l in days,j in workstations)
            Y[m][j][k][l] <= A[m][j];

    }

    And my .dat file, It connected data with Excel.

     

    Questions

    1. I want to know how to read output easily, maybe write the outputs to .txt file but I don't know how to use "lloOplOutplutFile" in cplex (Ofile) Could you all please give me some advice?

    2. I want to define that "workstation j=1 have to work along with workstation j=2 in same machine.", "workstation j=3 have to work along with workstation j=4 in same machine.", ... untill "workstation j=11 have to work along with workstation j=12 in same machine." How can I write ontraint in program??

    Thank you so much everyone.

    Kylie gp

     


    #DecisionOptimization
    #OPLusingCPLEXOptimizer