public static void Test()
{
var data = new FlexibleJobShopData("Test");
var j1 = data.MakeJob();
var j2 = data.MakeJob();
var m1 = data.MakeMachine();
var m2 = data.MakeMachine();
var m3 = data.MakeMachine();
j1.AddTask(
new List <Machine>()
{
m1, m2
},
new List <uint>()
{
3, 6
});
j1.AddTask(
new List <Machine>()
{
m1, m2
},
new List <uint>()
{
3, 3
});
j2.AddTask(
new List <Machine>()
{
m1, m2
},
new List <uint>()
{
3, 6
});
j2.AddTask(
new List <Machine>()
{
m2, m3
},
new List <uint>()
{
3, 2
});
j2.AddTask(
new List <Machine>()
{
m3
},
new List <uint>()
{
3
});
Solve(data, true);
}
public static void Solve(FlexibleJobShopData data, bool chatty = false)
{
// Compute horizon
var horizon = data.JobList.Sum(
j => j.TaskList.Sum(
t => t.Durations.Max()));
// ----- Create all intervals and vars -----
/*
* // Use some profiling and change the default parameters of the solver
* SolverParameters solverParams = new SolverParameters();
* // Change the profile level
* solverParams.profile_level = SolverParameters.NORMAL_PROFILING;
* // Constraint programming engine
* Solver solver = new Solver("JobShop", solverParams);
*/
// Constraint programming engine
Solver solver = new Solver($"FlexibleJobShop: {data.Name}");
// Initialize dictionaries to hold references to task intervals
var tasksByJobId = new Dictionary <uint, List <TaskAlternative> >();
foreach (var job in data.JobList)
{
tasksByJobId[job.Id] = new List <TaskAlternative>(job.TaskList.Count);
}
var tasksByMachineId = new Dictionary <uint, IntervalVarVector>();
foreach (var machine in data.MachineList)
{
tasksByMachineId[machine.Id] = new IntervalVarVector();
}
// Creates all individual interval variables and collect in dictionaries
foreach (var job in data.JobList)
{
foreach (var task in job.TaskList)
{
var alternative = new TaskAlternative(job.Id);
tasksByJobId[job.Id].Add(alternative);
var activeVariables = new IntVarVector();
var hasAlternatives = task.AlternativesCount > 1;
for (int alt = 0; alt < task.AlternativesCount; alt++)
{
var machine = task.Machines[alt];
var duration = task.Durations[alt];
var name = $"{task.Name}; Alternative {alt}: {machine.Name}, Duration {duration}";
IntervalVar interval = solver.MakeFixedDurationIntervalVar(0, horizon, duration, hasAlternatives, name);
alternative.Add(interval);
tasksByMachineId[machine.Id].Add(interval);
if (hasAlternatives)
{
activeVariables.Add(interval.PerformedExpr().Var());
}
}
alternative.AlternativeVar = solver.MakeIntVar(0, task.AlternativesCount - 1, task.Name);
if (hasAlternatives)
{
solver.Add(solver.MakeMapDomain(alternative.AlternativeVar, activeVariables));
}
}
}
// ----- Create model -----
// Create precedences inside jobs
foreach (var taskAltList in tasksByJobId.Values)
{
for (int i = 0; i < taskAltList.Count - 1; i++)
{
TaskAlternative currentTaskAlt = taskAltList[i];
TaskAlternative nextTaskAlt = taskAltList[i + 1];
foreach (var alt1 in currentTaskAlt)
{
foreach (var alt2 in nextTaskAlt)
{
solver.Add(solver.MakeIntervalVarRelation(
alt2, Solver.STARTS_AFTER_END, alt1));
}
}
}
}
// Collect alternative variables.
IntVarVector alternativeVariableVec = new IntVarVector();
foreach (var taskAltList in tasksByJobId.Values)
{
foreach (var taskAlt in taskAltList)
{
if (!taskAlt.AlternativeVar.Bound())
{
alternativeVariableVec.Add(taskAlt.AlternativeVar);
}
}
}
// Add disjunctive constraints on unary resources, and create
// sequence variables. A sequence variable is a dedicated variable
// whose job is to sequence interval variables.
SequenceVarVector allSequences = new SequenceVarVector();
foreach (var machine in data.MachineList)
{
DisjunctiveConstraint disjCt = solver.MakeDisjunctiveConstraint(
tasksByMachineId[machine.Id], machine.Name);
solver.Add(disjCt);
allSequences.Add(disjCt.SequenceVar());
}
// Create array of end_times of jobs
IntVarVector endsVec = new IntVarVector();
foreach (var taskAltList in tasksByJobId.Values)
{
TaskAlternative lastTaskAlt = taskAltList.Last();
foreach (var alt in lastTaskAlt)
{
endsVec.Add(alt.SafeEndExpr(-1).Var());
}
}
// Objective: minimize the makespan (maximum end times of all tasks)
// of the problem.
IntVar objectiveVar = solver.MakeMax(endsVec).Var();
OptimizeVar objectiveMonitor = solver.MakeMinimize(objectiveVar, 1);
// ----- Search monitors and decision builder -----
// This decision builder will assign all alternative variables.
DecisionBuilder alternativePhase = solver.MakePhase(alternativeVariableVec, Solver.CHOOSE_MIN_SIZE, Solver.ASSIGN_MIN_VALUE);
// This decision builder will rank all tasks on all machines.
DecisionBuilder sequencePhase = solver.MakePhase(allSequences, Solver.SEQUENCE_DEFAULT);
// After the ranking of tasks, the schedule is still loose and any
// task can be postponed at will. But, because the problem is now a PERT
// (http://en.wikipedia.org/wiki/Program_Evaluation_and_Review_Technique),
// we can schedule each task at its earliest start time. This is
// conveniently done by fixing the objective variable to its
// minimum value.
DecisionBuilder objectivePhase = solver.MakePhase(objectiveVar, Solver.CHOOSE_FIRST_UNBOUND, Solver.ASSIGN_MIN_VALUE);
// The main decision builder (ranks all tasks, then fixes the
// objective_variable).
DecisionBuilder mainPhase = solver.Compose(alternativePhase, sequencePhase, objectivePhase);
// Search log
const int kLogFrequency = 1000000;
SearchMonitor searchLog = solver.MakeSearchLog(kLogFrequency, objectiveMonitor);
const long FLAGS_time_limit_in_ms = 1000 * 60 * 20;
SearchLimit limit = null;
if (FLAGS_time_limit_in_ms > 0)
{
limit = solver.MakeTimeLimit(FLAGS_time_limit_in_ms);
}
SolutionCollector collector = solver.MakeLastSolutionCollector();
collector.AddObjective(objectiveVar);
collector.Add(alternativeVariableVec);
collector.Add(allSequences);
foreach (var taskVec in tasksByMachineId.Values)
{
foreach (var task in taskVec)
{
collector.Add(task.StartExpr().Var());
}
}
// ----- Search -----
bool solutionFound = solver.Solve(mainPhase, searchLog, objectiveMonitor, limit, collector);
if (solutionFound)
{
// The index of the solution from the collector
const int SOLUTION_INDEX = 0;
Assignment solution = collector.Solution(SOLUTION_INDEX);
Console.WriteLine();
uint machineIdx = 0;
foreach (var seq in allSequences)
{
machineIdx++;
var taskSeq = collector.ForwardSequence(SOLUTION_INDEX, seq);
Console.WriteLine($"{seq.Name()}:");
Console.WriteLine(" Tasks: " + string.Join(", ", taskSeq));
//foreach (var taskIndex in storedSequence)
//{
// IntervalVar task = sequence.Interval(taskIndex);
// long startMin = solution.StartMin(task);
// long startMax = solution.StartMax(task);
// if (startMin == startMax)
// {
// Console.WriteLine($"Task {task.Name()} starts at {startMin}.");
// }
// else
// {
// Console.WriteLine($"Task {task.Name()} starts between {startMin} and {startMax}.");
// }
//}
Console.WriteLine();
Console.Write(" Starting times:");
foreach (var s in taskSeq)
{
Console.Write(" " + collector.Value(0, tasksByMachineId[machineIdx][s].StartExpr().Var()).ToString());
}
Console.WriteLine();
}
//var x = tasksByMachineId[1][0].StartExpr().Var();
//var xValStr = collector.Value(0, x).ToString();
Console.WriteLine("objective function value = " + solution.ObjectiveValue()); //TEMP
}
// Save profile in file
//solver.ExportProfilingOverview("profile.txt");
// Done
solver.EndSearch();
}
public static SchedulingResults Solve(FlexibleJobShopData data)
{
}
