private AsynchSimulationResult ComputeSimulationResults()
{
// These double arrays store the double floating point result value for the simulation iteration at
// the index value. These distributions are then analyzed as probability models.
double[] completionTimes = new double[this.IterationCount];
double[] resourceCosts = new double[this.IterationCount];
double[] fixedCosts = new double[this.IterationCount];
// The resourceTime dictionary has the employee names as the key and the time usage distribution
// as the value.
var resourceTime = _workingProject.Organization.Employees.ToDictionary(organizationEmployee => organizationEmployee.Name, organizationEmployee => new double[IterationCount]);
// The taskStartTimes dictionary has the task ID as the key and the starting time distribution as the value
var taskStartTimes = _workingProject.Tasks.ToDictionary(x => x.Id, x => new double[IterationCount]);
// The taskEndTimes dictionary has the task ID as the key and the task ending time distribution as the value
var taskEndTimes = _workingProject.Tasks.ToDictionary(x => x.Id, x => new double[IterationCount]);
// The fixedCostsByCategory dictionary has the fixed cost category name as the key and the category total cost
// distribution as the value
var categories = new HashSet <string>(_workingProject.FixedCosts.Select(x => x.Category));
var fixedCostsByCategory = categories.ToDictionary(category => category, category => new double[this.IterationCount]);
var options = new ParallelOptions();
options.MaxDegreeOfParallelism = -1;
Parallel.For(0, IterationCount, options, i =>
{
var simulator = new ProjectSimulator(_workingProject);
var result = simulator.Simulate();
completionTimes[i] = result.TotalCompletionDays;
resourceCosts[i] = result.TotalResourceCost();
// Aggregate the usage times for each resource, by name
foreach (string resourceName in result.ResourceUtilization.Keys)
{
resourceTime[resourceName][i] = result.ResourceUtilization[resourceName].Select(x => x.Amount).Sum();
}
// Aggregate the distribution of start and end times for each task
foreach (var taskId in result.TaskStartTime.Keys)
{
taskStartTimes[taskId][i] = result.TaskStartTime[taskId];
taskEndTimes[taskId][i] = result.TaskEndTime[taskId];
}
// Aggregate the fixed costs
double totalFixedCost = 0;
foreach (string category in categories)
{
// Get the keys of the fixed costs in this category
var keys = _workingProject.FixedCosts.Where(x => x.Category == category).Select(x => x.Id).ToList();
// Get the total cost of these keys
double categoryCost = keys.Select(x => result.FixedCostValues[x]).Sum();
totalFixedCost += categoryCost;
fixedCostsByCategory[category][i] = categoryCost;
}
fixedCosts[i] = totalFixedCost;
});
var resultDistributions = new List <ProbabilityDensityData>();
var taskSimulationData = new Dictionary <Guid, TaskSimulationData>();
// Lambda to make sure that the result distribution is only added if not empty
var addResultDistribution = new Action <ProbabilityDensityData>(x =>
{
// if (!x.IsEmpty)
resultDistributions.Add(x);
});
// Main simulation results
addResultDistribution(new ProbabilityDensityData(completionTimes, "Project Duration")
{
XLabel = "Days",
Category = "Main",
ValueFormatter = x => $"{x:N0} days"
});
addResultDistribution(new ProbabilityDensityData(resourceCosts, "Resource Costs")
{
XLabel = "Dollars",
Category = "Main",
ValueFormatter = x => (x > 1000) ? $"${x / 1000:N0}k" : x.ToString("C0")
});
addResultDistribution(new ProbabilityDensityData(fixedCosts, "Fixed Project Costs")
{
XLabel = "Dollars",
Category = "Main",
ValueFormatter = x => (x > 1000) ? $"${x / 1000:N0}k" : x.ToString("C0")
});
// Resource Committment results
foreach (var resourceTimeData in resourceTime)
{
addResultDistribution(new ProbabilityDensityData(resourceTimeData.Value, $"{resourceTimeData.Key}")
{
Category = "Resource Commitment",
XLabel = "Days",
ValueFormatter = x => $"{x:N0} days"
});
}
// Fixed cost category
foreach (var fixedCost in fixedCostsByCategory)
{
addResultDistribution(new ProbabilityDensityData(fixedCost.Value, $"{fixedCost.Key}")
{
XLabel = "Dollars",
Category = "Project Fixed Costs",
ValueFormatter = x => (x > 1000) ? $"${x / 1000:N0}k" : x.ToString("C0")
});
}
// Task start and end times
foreach (var taskStart in taskStartTimes)
{
var task = _workingProject.GetTaskById(taskStart.Key);
var data = new ProbabilityDensityData(taskStart.Value, task.Name)
{
Category = "Task Start Day",
XLabel = "Days",
ValueFormatter = x => $"Day {x:N0}"
};
addResultDistribution(data);
taskSimulationData.Add(task.Id, new TaskSimulationData
{
MedianStart = data.MedianValue,
HighStart = data.UpperConfidence,
LowStart = data.LowerConfidence
});
}
foreach (var taskStart in taskEndTimes)
{
var task = _workingProject.GetTaskById(taskStart.Key);
var data = new ProbabilityDensityData(taskStart.Value, task.Name)
{
Category = "Task End Day",
XLabel = "Days",
ValueFormatter = x => $"Day {x:N0}"
};
addResultDistribution(data);
taskSimulationData[task.Id].MedianEnd = data.MedianValue;
taskSimulationData[task.Id].HighEnd = data.UpperConfidence;
taskSimulationData[task.Id].LowEnd = data.LowerConfidence;
}
return(new AsynchSimulationResult
{
Distributions = resultDistributions,
TaskData = taskSimulationData
});
}
