Given the alternative resource and processing time data generated above, the only difference in modeling between the standard job shop model and this problem is the declaration of the activities and the resource constraints.

Rather than having a constant processing time, the activities in this problem have an upper and lower bound. We, therefore, create each activity with a Solver variable representing its processing time.

      IloNum ptMin = IloMin(durations[k], altDurations[k]);
      IloNum ptMax = IloMax(durations[k], altDurations[k]);
      IloNumVar ptVar(env, ptMin, ptMax, ILOINT);

      IloActivity activity(env, ptVar);

      IloResourceConstraint rc =
        activity.requires(altResSets[resourceNumbers[k]]);
      
      // SET THE DIFFERENT DURATIONS DEPENDING ON 
      // RESOURCE SELECTION 
      rc.setProcessingTimeMax(resources[resourceNumbers[k]],
			      durations[k]);
      rc.setProcessingTimeMin(resources[resourceNumbers[k]],
			      durations[k]);
 
      rc.setProcessingTimeMax(
             resources[altResourceNumbers[resourceNumbers[k]]],
             altDurations[k]);
      rc.setProcessingTimeMin(
             resources[altResourceNumbers[resourceNumbers[k]]],
             altDurations[k]);
 
      model.add(rc);