Ranging and Floating


Words on this page that have been underlined and italicized can be found in the glossary on the right-hand side of the page.

Duration Ranging

Duration ranging is a method for modeling uncertainty in activity durations. As opposed to a single duration estimate, a set of possible durations (optimistic, pessimistic, intermediate vaues, etc.) is defined. The activity's duration is then sampled from its range for every iteration in the simulation.

Level of efforts, WBS summaries, and hammocks cannot have their durations ranged.

Tip

Duration ranging should only take into account the uncertainty inherent in duration estimates, while any other external factors should be captured as risks in the Risk Register.

  1. Click the Ranging  driver.  

  2. Click the View button above the top list and choose Duration Ranging.

  3. Scroll to the desired activity.

  4. In the Include column, check the box.

  5. In the Distribution Type column, click to choose a desired distribution type.

  6. Enter a value in each of the required columns.

Required columns will display a Warning icon. If none of the required columns are visible, they must be turned on (see Customizing Columns). So long as any column contains a warning, the entire row will be colored red.

Tip

At any time, rows with warnings can be found by typing "error" or "warning" into the search box above the list.

Tip

Turn on the Unit column to control whether the ranges to be entered will be in days or in percent variance. To turn on the Unit column, see Customizing Columns.

Sampling of In-Progress Durations

When activities are in-progress, the sampling behavior can be specified. 

  1. Click the Ranging  driver.  

  2. Click the View button above the top list and choose Duration Ranging.

  3. Scroll to the desired activity.

  4. In the Sampling Behavior column, click to choose a behavior.

Tip

To turn on the Sampling Behavior column, see Customizing Columns.

The following behaviors are available:

Sampling BehaviorDefinitionExample
Prorated

Default. The duration range is first prorated by the percent remaining and then added to the completed duration.

A 12 day activity with 3 days completed, 9 days remaining (75%), and a range from 8-16 days would instead be sampled from 6-12 days (75% of 8-16) and then added to 3 for a total duration of 9-15 days.
Adjusted then Sampled

If the completed duration is greater than the min, the min is adjusted; else, the range is sampled as-is.

A 12 day activity with 10 days completed, 2 days remaining, and a range from 8-16 days would instead be sampled from 11-16 days (since 10 have already elapsed). If only 7 days were completed, the range would continue to be from 8-16 days. In either case, the competed portion is not factored in.
As Sampled

The duration range is not changed but rather sampled as-is and then added to the completed duration.

A 12 day activity with 3 days completed, 9 days remaining, and a range from 8-16 days would continue to be sampled from 8-16 days and then added to 3 for a total duration of 11-19 days.

Floating and Pacing

Traditional schedule simulation places all activities on early starts, ignoring the impact of delayed or late starts (impact of float use) on project completion. However:

  • In the real-world, activities off the critical path are often started later than their early dates

  • This can be to level resources, pace progress, or for other strategic reasons

  • With traditional schedule simulation, probabilities of completion are often overestimated

Unique to NetRisk and the Graphical Path Method (GPM), non-critical activities can be made to start on their late dates through a process called floating.

  • Floated activities may end up on the longest path whereas otherwise they would not have

  • Combined with duration uncertainty and risk impacts, the probability of completion for the project may be pushed out

  • A true probability of completion can be inferred as somewhere in between the early curve and the late curve

As a rule of thumb, once a simulation has been ran and probabilistic completion dates have been determined, an additional simulation should be ran with all non-critical activities on late dates.

The following fields are available:

ColumnDescription
% of iterations

The percentage of iterations that floating will actually be applied.


To determine the late date distribution curve, set to 100%.

Float Threshold

An extra check above and beyond the % of iterations for whether or not to float an activity, since a delay in another path may increase available float during simulation. For example, a threshold of 150% requires the simulated float to be at least 1.5 times its deterministic float.

Float Threshold (continued)

A threshold of 50% requires the simulated float to be at least half of the deterministic float. A blank or empty threshold removes this extra check. When specified, the activity is being paced.


To determine the late date distribution curve, leave blank.

Distribution Type

A shape for determining the percentage of simulated float (the amount of float at the time the activity comes up for sampling) that should be consumed each iteration.


To determine the late date distribution curve, use Constant and set the Most Likely to 100%.

Delays cannot be floated and will not show up in the list.

  1. Click the Ranging  driver.  

  2. Click the View button above the top list and choose Floating.

  3. Scroll to the desired activity.

  4. In the Include column, check the box.

  5. In the Distribution Type column, click to choose a desired distribution type.

  6. Enter a value in each of the required columns.

Required columns will display a Warning icon. If none of the required columns are visible, they must be turned on (see Customizing Columns). So long as any column contains a warning, the entire row will be colored red.

Removing Sampling

The sampling of an activity's duration or float can be excluded from any simulation by unchecking the box in its row in the Include column. Even so, the distribution values will still be stored for future use.

  1. Click the Ranging  driver.  

  2. Scroll to the desired activity.

  3. In the Include column, uncheck the box.

Choosing Distribution Types

Choosing a distribution type should be determined by the level of confidence in the values being chosen:

  • The minimum is interpreted as the optimistic estimate, or best case scenario

  • The most likely is interpreted to be the most likely estimate

  • The maximum is interpreted as the pessimistic, or worst case scenario
Distribution TypeWhen to Use...
Constant
  • Use if there is no possible range, but rather only one possible value
Beta-Pert
  • Smoother than a triangular
  • Use if you are fairly confident in your estimate for the most likely but want to control how confident you are in the extremes
  • To use Modified Beta, set lambda to any value other than 4 (default)
  • If you set lambda to 0, the probability of the extremes will be lower than with, for example, a Triangular distribution
Log-Normal
  • Use if you know the duration is the product of several other variables
  • For example, many natural biological events can be modeled logarithmically, as can many variables in economics and finance
Normal
  • Use if the minimum estimate is just as likely to occur as the maximum estimate
  • Not recommended for durations, since an activity will be delayed more often than completed early
  • More typical with risk impacts
Triangular
  • Use if your confidence in the most likely is neither high nor low, as the probability of the extremes will be higher than a Beta-Pert but lower than a Uniform
Trigen
  • Use if what's been entered for the minimum and maximum may actually be too conservative, or if the minimum and maximum may not be the absolute minimum and maximum possible
  • Values below the Low % and above the High % will not be sampled
Uniform
  • Use if your confidence in the range is fairly low, as the probability will be the same for all possible values.
  • Not recommended for durations, since an activity will be delayed more often than completed early

Risk Assignments

Risks can be assigned via the lower table. For more info, see Risk-Activity Impacts.

Copy/Paste/Fill Down

Once entered, duration ranging, floating, or pacing values can be copied and pasted or filled down to save time. For details, see Interface Tour.

Available Table Columns

Some columns are unique to duration ranging whereas others are unique to floating. 

For details on customizing columns, filtering, sorting, and organizing by WBS, see Interface Tour.
 Ranging Columns...
ColumnDescription
% of IterationsFloating only. The percentage of iterations that floating will actually be applied.

Abbr. Description

A NetPoint-unique field limited to 48 characters for describing the activity.

Calendar

The schedule that work on the activity can take place.
Codes

A submenu containing one column for every code in the schedule.

The organization and visibility of code columns is NOT saved with the application. Code columns must be turned back on every time the tab is changed.

Completed DurationThe amount of time already completed.

Constraint #1

The primary constraint, if any. The following abbreviations are used: ALAP (As Late As Possible), MEO (Finish On), MEOA (Finish On Or After), FNL (Finish On Or Before), MS (Mandatory Start), MF (Mandatory Finish), MSO (Start On), SNE (Start On Or After), and MEOB (Start On Or Before).

Constraint #2

The secondary constraint, if any. The following abbreviations are used: ALAP (As Late As Possible), MEO (Finish On), MEOA (Finish On Or After), FNL (Finish On Or Before), MS (Mandatory Start), MF (Mandatory Finish), MSO (Start On), SNE (Start On Or After), and MEOB (Start On Or Before).
Constraint Behavior

A toggle for controlling how constraints behaves during simulation (only enabled if an activity/milestone has a constraint).

If an activity has two constraints and Maintain is chosen, both will be maintained unless one is triggered, in which case the second one will be ignored.

ALAP constraints will never be maintained and this option disabled for activities with only ALAP constraints (if an ALAP and another constraint exist and Maintain is chosen, only the other constraint will be maintained).

Description

The text field for describing the activity.

This column cannot be removed from the table when organized by WBS.

Distribution TypeA shape for organizing a range of values (e.g. durations, impacts, etc.).

Drift

The amount of the time an activity may advance before advancing the project start.

Duration

The amount of time the activity takes to complete.
Duration CorrelationsA summary of activities that have been correlated by ID with the coefficient showing in parenthesis.
Early FinishThe date the activity completes when its drift is equal to zero.
Early Start

The date the activity begins when its drift is equal to zero.

Existence ProbabilityDuration ranging only. The percent of iterations an activity will occur. When it occurs, the activity's duration is sampled from its range. When it doesn't occur, its duration is zeroed out for that iteration.

Finish Date

The date the activity completes. If actualized, the date is appended with an asterisk ( * ).
FloatThe amount of the time an activity may be delayed before delaying the project completion.
Float ThresholdFloating only. An extra check above and beyond the % of iterations for whether or not to float an activity, since a delay in another path may increase available float during simulation. For example, a threshold of 150% requires the simulated float to be at least 1.5 times its deterministic float. A threshold of 50% requires the simulated float to be at least half of the deterministic float. A blank or empty threshold removes this extra check. When specified, the activity is being paced.
High %A range value used when Trigen is chosen for the distribution type.
IDThe user-entered unique identifier.
Include

A checkbox for enabling or disabling the selected functionality.

This column cannot be removed from the table when organized by WBS.

LambdaA range value used when BetaPert is chosen for the distribution type.
Late FinishThe date the activity completes when its float is equal to zero.
Late StartThe date the activity begins when its float is equal to zero.
MaxA range value used for the following distribution types: BetaPert, Triangular, Trigen, and Uniform.
MeanA range value used for the following distribution types: Normal and Log-Normal.
MinA range value used for the following distribution types: BetaPert, Triangular, Trigen, and Uniform.
Most LikelyA range value used for the following distribution types: Constant, BetaPert, Triangular, Trigen, and Uniform.
Remaining DurationThe total duration minus the amount of time already completed.
Risk AssignmentsA summary of all risks and/or scenarios assigned by ID.
Sampling BehaviorDuration ranging only. The behavior for sampling of in-progress activities. For definitions and examples, see Sampling of In-Progress Durations above.
Sampling ValuesA summary for displaying the distribution type plus any range inputs together in the same field.
Start DateThe date the activity begins. If actualized, the date is appended with an asterisk ( * ).
Std.Dev.A range value used for the following distribution types: Normal and Log-Normal.
Total CostThe total cost of the activity, as calculated by resource assignments.
Total FloatThe float + drift.
TypeThe type of object, e.g. activity, milestone, hammock, etc.
UnitA toggle for ranging in days or % variance.
WBS SequenceA value for sorting when organized by WBS.
Jump to:


Still Need Help?

Click the icon above or visit our support desk.



 Glossary

Deterministic: a value from the base case schedule, i.e. non-simulated, for example for a duration, date, schedule, or other value.

Distribution: a range of possible outcomes (e.g. durations, impacts, etc.) together with a shape, from which a value will be sampled.

Distribution Type: a shape for organizing a range of possible outcomes (e.g. durations, impacts, etc.).

Floating: a technique that involves delaying non-critical activities as the simulation progresses to model resource leveling, progress-pacing, or other scenarios that come up in the real-world.

Iteration: one complete pass, from the beginning of the schedule to the end, whereby sampling of entered ranges is applied). All iterations taken together comprise a simulation.

Pacing: a technique that involves delaying non-critical activities as the simulation progresses, but only if the float when the activity comes up for sampling divided by the deterministic float exceeds the ratio specified by the Float Threshold.

Percent Variance: The percent difference (positive or negative) between a specified value and a fixed value, usually the deterministic duration. For example, a 0% variance corresponds to no change from the deterministic duration; a 100% variance corresponds to twice the deterministic duration; and -50% variance corresponds to half of the deterministic duration.

Simulation: a collection of one or more iterations taken together to represent or model a series of outcomes.