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.

Resource Leveling and "Optimization"

10 replies [Last post]
Brian Ultican
User offline. Last seen 16 years 40 weeks ago. Offline
Joined: 10 Jun 2007
Posts: 24
Groups: None
Im trying to familiarlize myself with how resource leveling works in the various scheduling packages and wanted to find out what the members here think about these features. I know how Microsoft Project does it but would like to find out how it is done in Primavera, Workbench, OpenPlan and Planview.

Everybody says that the way that Project does it is ’the worst’ but nobody ever says why the others are better.

Thanks in advance.

Replies

Rafael Davila
User offline. Last seen 20 hours 20 min ago. Offline
Joined: 1 Mar 2004
Posts: 5228

Merely hitting the resource leveling key is not enough to make your software resource leveling to be optimal or near optimal as true optimal might be missed or in complicated schedules you will never know. Near optimal at the present state of knowledge is enough, unfortunately operations research scientist have not reached to a 100% optimal algoritm. The best you can do is get software that consistently yields better results, and this you shall look for.

From http://www.stottlerhenke.com/products/aurora/index.htm

Most systems use simple rules to select and schedule activities and assign resources to carry them out. Often, these schedules are far from optimal.

My solution to the sample schedule provided in the following reference; http://www.stottlerhenke.com/products/aurora/Turnaround/2009-10-01_Auror... is as follows.

Photobucket

A 7 days schedule, obviously optimum solution as the unleveled schedule also takes 7 days. You should try it yourself.

Make sure your software resource leveling engine is good enough not only to provide better "optimal" schedules but also able to provide good prioritization control at different levels.

About mixing manual resource leveling with automatic resource leveling it is not considered good practice. For a reference on this issue click the following link.

http://www.ronwinterconsulting.com/Implementing_Resource_Leveling.pdf

if you are also going to later resource level your schedule ......

    Finally, there must be no resource considerations in the logic used. This is sometimes called, “soft logic.” You do not want to codify resource timing considerations into your schedule if you are also going to later resource level your schedule. The two different procedures used together will only produce erroneous results.

If resource-leveled schedules make most tasks critical or near-critical (using up most of the float in the process) then your resources are loaded in most effective way (idle time is minimized).

We already discussed manual leveling. It could be fine but it can be confirmed only by comparing manual schedule with the optimized schedule created by the software tool. And manual rescheduling after entering actual data is time consuming.

When we first time applied resource optimization in turnaround project at refinery plant and the project was successfully finished, we asked project manager if he saw advantages in using our software for resource management. An answer was "Yes, certainly! It was the first time when everything was done in available time window."  If you are disappointed with software application for project resource management then it may be two options: 1) software was not applied properly, 2) you used wrong software.

Dick Ertl
User offline. Last seen 11 years 39 weeks ago. Offline
Joined: 26 Jun 2012
Posts: 3

We prefer to set manpower hiring levels based on a well-sequenced and priritized schedule, which has been prepared with field supervisors' input.  This results in a workable schedule that field forces can accept.

 

Most resource-leveled schedules result in unworkable schedules since they make most tasks critical or near-critical (using up most of the float in the process).  They also result in a disruptive work flow, just to accomodate resource usage.

 

We use a different approach, that has proven practical and effective in countless turnarounds.  We notice that most resource-leveled schedules we have seen end up being used mainly for reporting to management; they are almost never used to help the field forces organize their work.  So field forces will complete the turnaround without schedules, just not very efficiently.

Could this be one of the elements that result in failed turnarounds?  All who have polled the industry agree that about 70% or more turnarounds end up in failure (extended schedules, huge cost overruns).

Rafael Davila
User offline. Last seen 20 hours 20 min ago. Offline
Joined: 1 Mar 2004
Posts: 5228

Applying priorities is fine if you can compare prioritized schedule with the optimal.

Of course, otherwise you are in limbo.

  • How do you know how good is your solution if you cannot compare to others that yield optimal or near optimal?
  • How do you know if it makes sense to keep the priorities of if it is time to make some adjustments?

Diick,

manual leveling does not guarantee good results, the software can estimate much more options but requires good project model where all schedule constraints are entered.

Besides, the situations are changing and the order of activities that was optimal yesterday may be far from optimal today.

Schedule recalculation is easy with the software resource leveling but manual resequencing of future works each time when actual data are entered is impossible. Applying priorities is fine if you can compare prioritized schedule with the optimal.

Rafael Davila
User offline. Last seen 20 hours 20 min ago. Offline
Joined: 1 Mar 2004
Posts: 5228

Usually at first what-if with software that allows for Phase Priority to take precedence over Activity Priority will give you the control you need.

If your urge for certain order is BIG then do not use optimization methods that might override some priorities in search for the optimum defined as of shorter duration, consider using Standard algorithms.

As you can see optimal is not the same for everyone and you shall select among the several algorithms available.

Look at the following schedule. All activities can occur at the same time, there are no links, only resource constraints. All activities sharing a same resource but higher priorities at phase level take precedence, then activity level will follow. Higher numeric value meaning higher priority.

Photobucket

Just imagine trying to do this manually in construction jobs where you frequently have partial assignments of multiple crews on each activity and shared equipment at different ratios. Like for example 100% of time formwork crew, 25% Electrical and Plumbing crews plus crane 50% of the day and this different on hundreds of activities.

I play Pampers Little Leagues, 700 activities per job is my average, still too complicated for ancient manual adjustments.

Dick Ertl
User offline. Last seen 11 years 39 weeks ago. Offline
Joined: 26 Jun 2012
Posts: 3

One problem with "automatic" resource leveling is that it generally makes most tasks critical or near-critical, which may be confusing to the field supervisors.  Additionally, if tasks are not sequenced logically, then the schedule may result in a disruptive work flow, with extra movement (mobilization/demobilization from one work item to the next, sometimes in unrelated parts of a process plant).

The best alternative is to sequence the work, in accordance with field supervisor's input and guidance, then prioritize the work.  This will result in a better schedule for process plant shutdowns/turnarounds.

 

Steve O
User offline. Last seen 3 years 30 weeks ago. Offline
Joined: 17 Jun 2012
Posts: -4

Hi,

Does anyone have an idea of how to optimize resource use without using leveling?

Hi Brian!

I did not try Workbench and Planview for quite a long time, so I will not discuss these packages and hope that somebody else will add an information on their resource levelling capabilities.

Microsoft Project has one single criterion for resource levelling (if not to count options like using or not activity priorities). This algorithm is not described - what you may find in MS Project Help describes different algorithm that was used before 2000.
So you can influence project schedule only manually using priorities and due dates. For large projects it is very hard to improve the schedule manually.

In Primavera and Open Plan you can select different priorities for resource assignments and through what if analysis find the solution that produce better results for current project.

In Spider Project there is an additional option that is called Optimization. If this option is selected then the software tries to improve created resource constrained schedule through iterative process using different criteria and their combinations. The schedule may become much shorter.

Besides different software can take into account different sets of constraints. For an example: project scheduling that considers financial and material supply constraints is supported only by Spider Project and partially by Open Plan (in this package you can set expense limits for time periods).

Regards,
Vladimir