Project(s) Prediction

In Epicflow, Prediction plans the order in which tasks are executed, based on their priority and business value, and on the real capacity of your resources. The system calculates task dates and shows them on the Gantt chart, so you can simulate project execution and resource allocation across the portfolio.

You can run Prediction for all active projects in the Pipeline, or for a single project.

Permissions

• Manage prediction – to run Prediction for the Pipeline.
• Manage single project prediction – to run Prediction for a single project.

Both are in Settings → Roles & Permissions → Project permissions. More about permissions in this article.

Prediction for all projects (Pipeline)

1. Open the Pipeline.
2. Click the planning type selector in the top-left corner (Screen #1).
3. Select Prediction to see the results of the last Prediction run.
4. To run a new calculation, click the recalculate icon (↻) next to Prediction. Hover over the icon to see who ran the last Prediction and when.
5. The Pipeline prediction window opens (Screen #2). Set the parameters (see “Prediction parameters” below).
6. Click Run.

Screen #1 – Planning types

Screen #2 – Pipeline prediction

Prediction for a single project

Single project prediction works the same way, but only for the opened project. Use it to plan and adjust one project within its own constraints.

1. Open the project in the Pipeline (Project editing / Quick Edit Mode).
2. Click the planning type selector and select Run Prediction (single project).
3. The Single project prediction window opens (Screen #3). Set the parameters.
4. Click Run.

Screen #3 – Single project prediction

You can read about other planning types in this article.

Prediction parameters

• Scope for prediction (Pipeline only):
  • All active projects.
  • Filtered only – only the projects that match the filter applied in the Pipeline. This option is available only when a filter is applied.
• Configuration – a saved set of parameters. Open the list to select a configuration or click Save new. Hover over a configuration to rename, copy or delete it (Screen #4). The configuration in use is marked Used. When you change parameters, the configuration is marked (edited).
• Parameter impact on the task order – a slider with seven positions, from Task priority to Task business value. Hover over a position to see its name:
  • Completely priority oriented (100/0)
  • Significant impact of priority (80/20)
  • Impact of priority (60/40)
  • Balanced (50/50)
  • Impact of business value (40/60)
  • Significant impact of business value (20/80)
  • Completely business value oriented (0/100)
• Allow temporary multitasking – temporarily switch to shorter tasks when feasible.
• Continue prediction after failure – if a task can’t be predicted, prediction continues for its successors. The failed task is skipped with a minimal duration.
• Intermediate milestones block next tasks – next tasks can’t start earlier than preceding intermediate milestones.

Screen #4 – Configurations

Advanced prediction

Advanced mode lets you set prediction parameters per group or resource.

1. In the Prediction window, turn on the Advanced toggle in the top-right corner (Screen #5).
2. In the resource table, set Prediction approach and In available slack for All resources or for a specific group or resource. Use Search resource and Filter by prediction approach to find them.
3. Click Save & Run. To keep the changed parameters as a separate configuration, click Save new next to Configuration first.

Screen #5 – Advanced prediction

Prediction approach

Click the value in the Prediction approach column and select (Screen #6):

• User in group capacity – the system assigns specific users to tasks. Pre-assigned users are always considered first.
• Group capacity – the system ignores specific user assignments and uses the overall capacity of the group. Users who work in several groups are taken into account to avoid underutilization.
• Infinite Capacity – capacity does not limit the calculation.

Screen #6 – Prediction approach

In available slack

Defines when a task is started even if no suitable resource is available. Click the value to switch between without and within. With within, set the priority threshold (Screen #7):

• 100 – the task starts just in time to be completed before the milestone.
• 80 – the task starts when its priority reaches 80 or more.

Screen #7 – In available slack

Possible usage scenarios

• Identify bottlenecks within a group.
• Override group capacity for specific resources right before a Prediction run, instead of changing capacity in the resource settings.
• Simulate the impact of adding resources (hiring or training) without changing the Resource Pool.

Prediction-driven delays

A prediction delay is the difference between when a task could start and when it is predicted to start. To analyse it, compare the Possible dates (from the Infinite capacity plan) with the Prediction data.

In the Group Task List

1. Run Prediction in the Pipeline.
2. Open Task List → Group and select Prediction in the planning type selector.
3. Click the gear icon in the top-right corner.
4. Under Sections, select Timeline.
5. Under Timeline settings, select Possible dates and Prediction data (RTS) (Screen #8).

Screen #8 – Enable prediction data in the Group Task List

Prediction data (RTS) shows where the task would be (RTS), relative to its earliest start date, if only the delay of its predecessors were considered and resource delay were ignored. The difference between the RTS bar and the predicted bar is the resource delay. In the Infinite capacity plan, this option shows the prediction bar.

On the task card

Open the task card and find the Prediction block (Screen #9). It shows the Estimated start, Estimated finish and Resourcing delay – the delay caused by resource availability.

Screen #9 – Prediction on the task card

Example

In the example below, the ‘Logistic Engineering’ task is not delayed: its prediction dates stay within the possible date range. The ‘Mechanical Assembly’ task has a prediction delay: its prediction dates go beyond the possible date range.

Screen #10 – Prediction delay example

System attribute “Prioritized” for tasks

You can mark specific tasks as Prioritized for leveling. Such a task:

• gets a higher priority in leveling and Prediction runs;
• gets this priority regardless of other indicators, such as task priority, business value or default scheduling rules;
• is scheduled and resourced earlier than other tasks.
• Screen #11 – System attribute “Prioritized” for tasks

Use it for critical contractual obligations, regulatory or compliance tasks, and other time-sensitive work.