Scheduling Reality: AI and the Critical Path Method (CPM)
April 15, 2021
Today, nearly every capital project uses the critical path method (CPM) for planning. This traditional method, however, has earned a less than stellar reputation for accuracy. This can leave projects failing not because of poor execution, but because the plan was likely overly optimistic or aggressive.
Fortunately, there is a better way. Thanks to emerging technologies, today’s planners and schedulers can deliver realistic project plans in a fraction of the traditional time by leaning on augmented intelligence (AI). AI helps provide smart suggestions based upon an organization’s historical project knowledge. AI and CPM can now work together, enhancing and complementing each other to yield more realistic schedules than ever before.
As one of the most detail-oriented scheduling techniques, CPM lends itself quite well to larger, more complex builds because it can account for the thousands of action items involved.
What is meant by a “critical path?” It’s a set of activities that must occur in a particular sequence in order to complete a project. This path, showing the steps before and after each task, illustrates the dependent relationships between tasks as a network or web, reflecting the duration of not only individual tasks but the overall project.
Start and finish dates and times are assigned to each task, allowing CPM’s algorithm to outline the critical path. The beauty of this method is it provides insight into the impact of a single schedule change on all dependent tasks.
With the critical path mapped out, project managers can determine how resources — such as labor or equipment — should be distributed for the highest level of efficiency.
How AI can Help
In just the past several years, the speed and increasing affordability of computing power has utterly changed the AI landscape, enabling the technology to reside on a planner’s desktop or in the cloud. From a planning perspective, that means a computer is not merely a mechanism for receiving inputs anymore but is actually proactively guiding the planner.
Though AI is not some magic tool whereby a contractor pushes a button and generates a schedule, it can be leveraged and manipulated within CPM to provide more realistic, dependable suggestions. In effect, it improves the planning process by enabling a computer to mimic the way a human makes decisions.
The proliferation of AI has helped fuel a knowledge-based approach to planning. For example, if a plan was originally created in Primavera P6® or Microsoft Project®, those schedules and plans will reside in a knowledge library that stores historical schedule information as well as unstructured data such as risks and opportunities.
The “brain” of the knowledge-based system is the inference engine. When a planner asks for a suggestion, the inference engine requests inputs in order to understand the context, such as the type of project, location, quantities and other variables. It then uses those variables to understand context, interrogate the knowledge library and provide suggestions.
Putting an End To “From-Scratch” Planning
With an augmented intelligence planning approach paired with CPM, you don’t have to begin with a blank sheet of paper anymore, but instead can rely upon historical data as a foundation.
By using a template, a planner can begin with some reliable building blocks for putting the plan together – and adjust along the way. For example, instead of having a work breakdown structure go 12 levels deep, a contractor can now reduce that number while retaining much of the data.
Today’s scheduling tools can also define high-level timelines, thereby providing a benchmark plan as the project progresses. Along the way, a contractor can better understand the alignment – or misalignment – of the plans, and whether or not they’re executing as per the original plan.
Task durations can also be established using historical rates of productivity. In the process, the AI-enhanced program asks a series of questions (e.g., scope of work, number of units, and so forth) and then uses historical productivity rates to determine the duration. This provides a more defendable, realistic estimate that is based upon past performance.
Through it all, the human component remains vital and necessary, as plan validation must be performed by the scheduler, team members and other stakeholders. The computer then takes those opinions and develops a consensus view.
With the incorporation of augmented intelligence planning capabilities and CPM, the entire planning process has become more of a collaborative partnership between computer and planner, offering a quantifiable leap forward in the world of project planning and scheduling.