Guild of Project Controls: Compendium | Roles | Assessment | Certifications | Membership

Tips on using this forum..

(1) Explain your problem, don't simply post "This isn't working". What were you doing when you faced the problem? What have you tried to resolve - did you look for a solution using "Search" ? Has it happened just once or several times?

(2) It's also good to get feedback when a solution is found, return to the original post to explain how it was resolved so that more people can also use the results.

Declining levels of knowledge about CPM

17 replies [Last post]
Patrick Weaver
User offline. Last seen 5 days 23 hours ago. Offline
Joined: 18 Jan 2001
Posts: 372
Groups: None

A sad and worrying trend appears to be the steadily declining levels of knowledge about CPM, and its origins ranging from a resent 'white paper' promoted by Planning Planet that confused PDM, ADM, CPM and PERT, through to blatent lies promoted by some 'lean construction' advocates claimg CPM cannot do resource analysis. 

A brief summary: 

CPM and PERT both used ADM when develoed in 1957,  PDM (Precedence) was made public in 1962 and was included in both CPM and PERT software shortly thereafter. 

CPM (not PERT) was developed to resolve resource issues and has progressively included more sophiticated resouce calculations, PERT did not consider resources until PERT/Cost in 1961. All modern CPM softare include reource calculations (they are't always used but that is a different issue). 

CPM is not a factual statement about what will occur in the future (this is impossible)

These topics are discussed in more detail in my latest blog post Critical Path Scheduling - 4 things people don't gethttps://mosaicprojects.wordpress.com/2022/07/27/critical-path-scheduling-4-things-people-dont-get/

Replies

Alex Lyaschenko
User offline. Last seen 1 week 3 days ago. Offline
Joined: 15 Aug 2011
Posts: 59
Groups: None

Patrick,

There are many tools that claim that they have "Resource levelling" but quality does matter!

 

In theory Primavera and MS Project also can level resources but experience schedulers know that this feature never should be applied. Both tools have so bad algorithms that it is better to find a solution manually.

 

I am using any chance to compare when I can. My solutions always were better than P6 and MSP could offer but not as good as calculated in Spider Project. I refer to simple cases with 1-2 team(s) and 20-30 tasks.

 

In real life when 100+ resources have different productivities and calendars smart levelling could significantly shorter the duration of projects. It is not 3-5%, but 20-30% or even 50% shorter!!! I can guarantee that $1B projects (or portfolios) can save at least 10% if they use a tool with smart levelling. It is $100M!

 

Unfortunately, we have steadily declining levels of knowledge in scheduling, not just CPM!!!

90% of schedulers I have interviewed know how to use a scheduling tool but never learn the scheduling. 

Santosh,

what is core CPM activity duration?

We enter volumes of work in physical units (meters, tons, etc.), assign resource crews, enter resource productivity and workload and (in simple case) get activity duration.

Examples of complications:

- assigning resource skills and the software will select who will do what in the process of scheduling. As the result we will learn activity duration because resources may have different productivity,

- assigning variable resource quantities (minimum and maximum) and once again we will learn activity duration after scheduling.

In any case activity duration does not depend on CPM. CPM is used for calculating project duration.

Santosh Bhat
User offline. Last seen 51 weeks 2 days ago. Offline
Joined: 15 Apr 2005
Posts: 381

Consider a simple example of laying a pipe, there are many ways this can be reflected in a schedule. Does the duration for the task include the mobilisation and demobilisation time? Or are these discrete activities for each event. Does the duration assume a single crew or multiple crew? Does the duration take into consideration the the access to complete the works? intereference from other works? the availability of materials and supply to the site. What about learning curves associated with work?

I consider resourcing as one of, potentially many inputs into determining durations of activities. Any method that takes the core CPM activity duration + logic, and then modifies this method to reflect all the above issues is a layer above CPM, and is then open for interpretation by the algorithms used (consider how each tool applied resource levelling).

So I think the term "Based on CPM" is probably more accurate. 

Rafael Davila
User offline. Last seen 11 hours 1 min ago. Offline
Joined: 1 Mar 2004
Posts: 5229

Software of the 1970's leveled effort but not resources, they got it wrong and after many decades still do not realize it. Leveling effort is not equivalent to leveling quantities and can yield impossible results.

Always level of knowledge of CPM have been so low that there is no room for declining.

David,

the first reason is the main.

Our experience showed that small and middle sized companies, where the bosses (and owners) are involved in project management and count money, people use resource leveling because they need realistic schedules and cost estimates.

In large companies bosses know better and serious planning is rare. They think that they must set targets and meeting targets is the responsibility of their subordinates, contractors, etc. Existing limitations and risks are not taken into account. As the result projects are late and finding as many reasons as possible why did it happen is the main art required from project managers.

I remember one case when I told top manager of telecom company that they will not be able to finish three projects of their portfolio before the end of the year. He answered that they do such projects in six month and I am wrong. At the end of the year the projects were not finished and I reminded that it was predicted. He answered that these projects were at special attention, he insisted to do project works faster and it helped to finish as Spider Project predicted. After this experience this company became much more serious about project planning and required resource, cost, supply and cost leveled schedules from its planners.

Top managers education and motivation is the main problem.

David Kelly
User offline. Last seen 1 year 34 weeks ago. Offline
Joined: 19 Oct 2004
Posts: 630

Well, this is a bit of fun!  A much more interesting thread than ‘Why is my bar red?’ Q&A.

In my not very humble opinion, the three most important questions that must be answered to manage a project are:

What is the scope of work?

What is the scope of work?

And

What is the scope of work?

 

A network diagram of so-called ‘activities’, with ‘relationships’ based on a simple understanding of the start and finish nodes for each activity do not help us understand the scope of work.

The Critical Path Method, calculating early and late dates and deriving float, allows us to draw a cute graphical representation of the schedule. This diagram is both useful and dangerous. It is dangerous if it is purely based on activity durations and not scope of work, aka resources.

I first used resource levelling of a CPM project diagram in 1976 – PSDI’s ‘Project 2’ did a very good job for the time. The ‘un-levelled’ network only took two weeks to complete but we would have needed two thousand men offshore, and only had accommodation for seven hundred. I fully expected all of the projects that I would be involved with from then on to make levelling essential. That has been true of less than 10% of my projects in the last fortyish years.

Why?

1)      The boss does not understand it.  The simple arithmetic of the forward and backward pass can be explained to management. Resource levelling algorithms are diverse and complex. When I wrote one for Data General mini computers forty years ago, I used a minimum moment matrix algebra algorithm which I extended to three dimensions so that you could rank resources for best fit. This would have been too difficult to explain to many project managers. BUT the engineering manager accepts FE calculations from the computer without reviewing the maths, so what is the problem here?

2)      The estimate of the scope of work is poor or even missing completely. This disaster in waiting is easily disguised, by not using resource levelling

3)      Primavera is simply not good enough. How many of you who chose Primavera investigated this before purchase?

4)      Very happy for you to extend this list!

 

 

 

 


 

Santosh,

I agree. But it means that resource leveling is not CPM scheduling even if it determines resource constrained floats and resource critical path (critical sequence).

Pat has different opinion.

It is necessary to come to common understanding of the term CPM scheduling to avoid unnecessary debates.

We need Definition.

Santosh Bhat
User offline. Last seen 51 weeks 2 days ago. Offline
Joined: 15 Apr 2005
Posts: 381

Valdimir, I don't think we should call any form of project scheduling as CPM. For example if you're only running a forward pass, then you cannot determine float, so there is no "critical path" determined by float values. However, there may be what we might call a "Driving Activity Path", as is used in Linear Scheduling methods (see this image for example: https://www.turbo-chart.com/wp-content/uploads/2019/12/linear-scheduling.jpg )

 

 

In my opinion (and I may very well be wrong about it) CPM does not concern itself with resources, or any other contraints unless these are implied in the durations and logic develoepd between tasks.

Adding Resource levelling, or any other automated process of developing logic and/or durations (for example auto-generating physical constraints such as Floor 1 before Floor2 etc) is a layer above the core CPM algorithms.

Patrick,

I know that resource leveling was and is used for a long time by different tools.

My first PM software developed in 1978 for mainframes included resource leveling also.

I like your presentation on resource optimization. When we launched Spider Project in 1993 I expected that people will be interested in its resource optimization capabilities. I made presentations on resource optimization, resource constrained floats, resource critical path in 1996 at PMI Conference and IPMA Congress and did not meet understanding. Now I know that few planners level their project resources and Spider Project attracts more new users by its many advanced functional features where resource constrained schedule optimization is not the main one. Resource optimization can bring huge benefits but people do not understand this or are not interested. But is it CPM?

I just asked to determine what is CPM.

Does it mean applying CPM algorithms or do you call CPM any kind of project scheduling?

We need generally accepted definition what is CPM. Can you suggest one?

Patrick Weaver
User offline. Last seen 5 days 23 hours ago. Offline
Joined: 18 Jan 2001
Posts: 372
Groups: None

Vladimir,

You are completely wrong - just read any book on CMP Resource levelling published since 1965.  If you are applying resource levelling to the CPM schedule activities are delayed if the resources are not availble to do the work, usually because they are allocated to work on other activities.  All CPM (and PERT) software since around 1965 has had this capability including mainframe systems developed by IBM, ICL, Burroughs Corp., Sperry Corp., and others.  I know I was using these tools from the early 1970s and still have access to some of the early manuals.  And we were doing resource levelling! 

The computaions were based on decision tables, rather than logic, the operators could adjust the priorities in the decision table and override the maths by including resource flows in the schedule logic.  ALL of these capabilities carry forward int tools such as Primavera, Microsoft Project and Microplanner. 

The limitations of this approach are based on the computig power of early 1960s computers and involved a lot of work - 60 years later there are better options focused on resource optimization and artificial intellegence (specifically machine learning) but perpetuating the lie that CPM cannot (or does not) do resource levelling is of no help to anyone looking to improve the practice of project controls.  Where we should be going is diccussed in: https://mosaicprojects.com.au/PMKI-SCH-013.php#Process5

Patrick Weaver
User offline. Last seen 5 days 23 hours ago. Offline
Joined: 18 Jan 2001
Posts: 372
Groups: None

The definition of the 'critical path' is highly varied (one of Vladimir's points), see:  https://www.mosaicprojects.com.au/WhitePapers/WP1043_Critical_Path.pdf

Then there's the subject of my post that started this chain - the historical facts around critical path scheduling (CPM), which incudes fixed durations, resource allocations and resource accregation, smoothing and levelling and the fact people routinely confuse smoothing and levelling (there's dozens of 'experts' get these the wrong way around). 

Then there's the spave Vlaidimir is already working in - resourec optimization where the focus of scheduling shifts from arbitary activities and durations (the core component of CPM) to focus on resource productivity and work sequence. This is an emerging space with very little support in contract law or commercial tools but should be the future of project controls. My first paper on this was published in 2011: https://mosaicprojects.com.au/PMKI-SCH-013.php#Process5

But there is still a very long way to go. 

Santosh,

I like it. By your definition Critical Path can consist of independent activities and is actually Critical Sequence (not path).

But I don't agree with "given durations" because they frequently depend on other parameters like activity volume of work and assigned resource productivity and are calculated in the process of scheduling.

 So CPM is any method that calculates Critical Sequence.

So if the software can calculate Resource Critical Path and resource constrained floats it is CPM software. But how we will call MSP and P6 that calculate wrong floats in resource constrained schedules and so do not determine Resource Constrained Critical Path? Do you agree that they are not CPM tools?

Santosh Bhat
User offline. Last seen 51 weeks 2 days ago. Offline
Joined: 15 Apr 2005
Posts: 381

Vladimir.

A method of sequencing tasks of given durations that uses the difference between a forward and backward calculation to determine float.

Hows that?

Santosh,

please answer a simple question: what is CPM?

Santosh Bhat
User offline. Last seen 51 weeks 2 days ago. Offline
Joined: 15 Apr 2005
Posts: 381

Pat,

I note your absence on the other platform - you are missed and it's a loss for users over there. But thankfully we know where to find you.

They took your information and thought they had found a silver bullet explaining why CPM is a failure. One perspective I’ve had is that they didn't appreciate the paper was published not long after CPM (as an algorithmic process) was developed, and the authors were not defining it as failure, but an area they acknowledged required further work, which is very common in all new areas of research and discovery.

My own perspective is that, yeah CPM doesn't acknowledge resources, nor does it acknowledge logical sequencing of work, how you're going to do the work, what durations to use or what relationships to use. It doesn't understand what you're doing, in fact it doesn't even care what you're using the calculations for - be it planning a holiday or managing a complex project.

Who seriously expects CPM to provide all the answers for them? It’s like asking a calculator to make important financial decisions for you.

My concern about these arguments is based on several things, including:

 

  1. That all these alternative techniques proposed to supersede CPM are typically proprietary methods, and the argument against CPM is more a marketing spin to convince others to use that registered, trademarked alternative on use-for-a fee basis - or simply, buy our stuff and your life will magically become better. Selling the dream!
  2. Many of the alternatives propose to make it easier, or simpler, with the conclusion that it makes these alternatives more accessible and therefore ensures greater collaboration, successful outcomes etc. BUT they are either simplifying the process to a point where it's not actually a meaningful tool for managing complex projects (ask if their technique can identify which activities are critical?), or if you drill deeper into the tools, they ultimately are just a variation of CPM anyway.
  3. Why should we try to simplify what is in essence a very complex operation to manage and deliver projects?  yes CPM might be difficult for some, but projects are difficult also.
  4. There was one claim that 90% of projects are delayed purely because they use Primavera as their scheduling too. Aside from not having evidence to back that claim up, it's "clickbait" statements like this, from "industry experts" that may mislead those who may not be as knowledgeable. I don’t believe there has ever been any solid research to understand if there is any correlation between results and the scheduling method adopted? I know the lean advocates imply there is.
  5. A lot of the lean advocates suggest their alternatives are about improved collaboration, reduced waste etc. I don't know how others feel, but I'm a bit insulted that this assumes all the work we do is wasteful and anti-collaborative etc. As I wrote elsewhere, is pragmatism exclusive to those who hole Lean Certifications? or do I owe years of royalties to someone who trademarked common sense?

and finally, and slightly tangential

  1. Many hope AI and Machine Learning are the solution. But I don't believe these will be anywhere near as useful as expected. Why? Schedules are rubbish data. They look fantastic from a data analytics perspective, thousands of rows of activities and relationships across millions of schedule data files. But what is the data you're analysing? I think all the AI/ML tools that drill into Schedule data are only going to clarify how schedules are built, not in how they deliver the projects they represent. To understand how those schedules were developed and what causes projects to be delayed, disasters or successful is not captured in schedule data. The data to drive those improvements is simply not in schedules. I think we have a mountain of improvements to make in the scheduling field to let AI/ML start using schedules, and CPM methods should still be the way this is achieved.

Of course all the above is my own opinion and I am open to being shown how wrong I am on any of the above.

Praveen Malik
User offline. Last seen 1 year 27 weeks ago. Offline
Joined: 20 Aug 2014
Posts: 13
Groups: None

Hi Patrick, Can you please point to the white paper you alluded to?

There are many misconceptions about Critical Path.

If you search on Internet, you will find many aticles claiming CPM and PERT as same. They were developed by different organizations and are altogether differnt things but they can be used together.

Here is a lowdown on how to use CPM and PERT together - https://www.pmbypm.com/critical-path-pert-and-standard-deviation/

And FAQ on Critical Path - https://www.pmbypm.com/critical-path-method-cpm-project-management/

Patrick,

Critical Path Method algorithm does not consider resource constraints.

Critical Path calculated taking into account resource constraints (Critical Sequence, Resource Critical Path) can consist of activities that are not linked with each other. It is calculated using different algorithms (usually resource leveling heuristics).

We need generally accepted definition what is CPM. Can you suggest one?