Graphical Path Method

Innovative algorithms for planning and scheduling

Innovative Vision

A compelling interview with Dr. Gui, the visionary CEO/Founder of PMA Consultants, as he explores the driving force behind the company’s commitment to innovation. Discover insights into PMA’s remarkable achievements, highlighted by the creation of the Graphical Path Method (GPM). This groundbreaking approach addresses inefficiencies in traditional methods, offering self-healing algorithms that enhance control over schedule activities. Notably, GPM and our software solutions have secured nine patents, with more to come. Dr. Gui articulates his vision for PMA Consultants—a company defined and differentiated by innovation. With this mindset, possibilities unfold, revealing new perspectives for a brighter future.

What makes GPM unique?

NetPoint Animated in Real Time

Real-time calculations

When an activity is moved within a network, its relational logic is subject to breaking down. GPM heals these relationships by propagating the change throughout the network in real-time. This automatic correction is known as ‘self-healing’ and the result is dynamic preservation of logic. Since GPM doesn’t rely on a database-driven scheduling engine – dates, durations, float attributes, the critical path, and resource profiles all continuously refresh as the network is modified. These real-time calculations provide invaluable information as schedules are built.

A network being updated as an activity is moved. Highlighted in blue: durations, link gaps, and total floats.

Planned Dates screenshot


GPM omits the early start bias inherent with CPM, permitting activities to be placed anywhere within their total float range, without constraining float. This is achieved with the introduction of Drift: the number of time units an activity can gain before advancing the start of the project. In a GPM network, Drift plus GPM Float is equivalent to CPM total float. With the ability to draw and stretch activities in either direction, planned dates allow projects to be planned forward or backward, allowing schedules to be built more naturally and with more flexibility than CPM.

Ext. Mason Walls shown in between early dates and late dates without a constraint. Highlighted in blue: link gaps, total floats, & drift and float.

Forensic Float screenshot

Forensic float

In CPM, the ability to calculate floats is lost when actual dates are entered. Not only does GPM retain this ability, but it’s inherent to the GPM algorithm. Once activities are actually completed, the meaning of float is interpreted as hypothetical: how much could an activity have been delayed? In GPM, this value is referred to as forensic float, and it continues to refresh even as schedule progress updates to the right of the data date. This can be particularly useful when analyzing a schedule for delays after the project has started or finished.

Joists & Desk completed with 1 day of forensic float. Highlighted in blue: link gaps, total floats, & forensic total float.

Scheduling method comparison

Critical Path MethodGraphical Path Method™
Diagramming Method Precedence Diagramming Method Logic Diagramming Method
Activity PlacementCPM Early DatesEarly, Late, or Planned Dates
Link Leeway Calculation None Gap
Activity Early DatesCalculated from Forward Pass Current Dates Less Drift
Activity Late Dates Calculated from Backward Pass Current Dates Plus Float
Activity Free FloatCPM Free FloatGPM® Buffer and Drift-Buffer
Can Activity Float Forward?If Activity Total Float is > ZeroIf Activity Float is > Zero
Can Activity Float Back?No, Activity Drift Does Not ExistIf Activity Drift is > Zero
Floating by Delaying StartReduces Total FloatDoes Not Reduce Total Float
Float Governing TechniqueThrough Mandatory ConstraintsBenchmarks as Fixed Events
Total Floats Left of Data DateCannot be Determined (Always = Zero)Recalculated if Schedule Changes
Software Data StructureDatabaseObjectbase
Scheduling Algorithm CPM Block Scheduling AlgorithmStep-wise Algorithm
Software User InterfaceBatch (Multiple Engines)Interactive, Real-time