Module 09 Spring 2006 closure

Module 9 Spring 2006 closure
Prepared by Greg Kinney

Most meaningful

The most meaningful I learned in this session was to get more into Microsoft Project, it was good to get some more knowledge about this program and all its possibilities.
The module gave a good understanding of the parts in modeling a network, for example how to determine Slack.

The Microsoft Project tutorials have been fantastic. The use of charts is an effective and easy-to-read method of presenting and estimating projects. The software is quite user friendly and has just a ton of options. My only problem is the firm I work for uses Suretrac. Instructor replies: I haven’t used Suretrak personally but the overwhelming consensus I’ve heard is that it is superior to Microsoft Project. So you’re probably fortunate!

The most useful and meaningful thing I learned in this module was the importance of having accurate information for making decisions regarding the critical path. In particular, the importance of variance and standard deviation for identifying the critical path within a network is paramount to the success of a project. The homework assignment for this module was both lengthy and intense. Nevertheless, it was this intensity that helped me to overcome the intimidation posed by both Microsoft Project and Crystal Ball. I felt very comfortable with both programs by the end of the homework assignment.

This chapter has provided much needed clarification on the differences between AON and AOA network diagrams and their appropriate use. I have seen both used interchangeably with incorrect notation. Activity-on-node diagrams are used to show the flow of time, from left to right, with the project tasks denoted on the rectangular nodes, whereas activity-on-arrow display the flow of work with milestones, or events, occurring at circular nodes and labeled arrows showing the tasks.

The simulation problem and probability calculations with Z–values were pretty straight forward.

I really enjoyed and found useful the use of Microsoft Project. I’ve never used it before and had fun tinkering with it.

Muddiest

Comment: "The muddiest" was all the different abbreviations in the different diagram types, this really took me a while to get into! (CPM, ADM, PERT, PDM)

Instructor reply: No question to reply to.

Comment: How to use Table 8.5 on page 399. It showed how to use the table when Z was positive. However, what if you are trying to find the probability of completing the project early. This would give you a negative Z. Therefore, can you use Table 8.5 for such values.

Instructor reply: Yes – this topic is covered in statistics textbooks but isn’t here. When you’re using a Z value in a normal probability distribution, you are interested in the cumulative probability of the distribution to the left of the z value. Let’s say there are two values, +z and –z, where –z is just the negative of +z. Then if P represents the single tail (left hand) probability that you find in Table 8-5, then the probability associated with –z turns out to be (1-P). For instance, at z = 0.20, P = 0.5793. For z = -0.20, P = (1-0.5793) = 0.4207. An easy way to calculate this – easier than using Table 8-5 – is to go into Excel and use the NORMSDIST function. All you do is input your value of z, positive or negative, and it spits out the answer.

There is much more to know about this and I really don’t recommend relying on Excel as a substitute for understanding this. If you can see your way clear to do it, AS 307/308 are terrific courses for understanding the fundamentals, and the Statistics for ESM course is an excellent applied course. The way engineering is evolving, we are all going to have to get much better at this.

Comment: The problem of choosing a particular method to perform quantitative risk analysis is still confusing to me. I understand the standard statistical methods described for uncertainty in project completion time. But I’m not following Crystal Ball very well at all and not quite sure how I might use it for construction estimating and project scheduling.

Instructor reply: A course in statistics, and then another course in statistics, will clear up a lot of the mystery around this. I recommend these strongly. Apart from that, part of the problem has to be the fact that Microsoft Project doesn’t handle simulations internal to the package. Some project management software now does. So you have to get out of Microsoft Project anad the get into the Crystal Ball package in order to get anywhere. This is confusing, and that part of it would be better just with some practice.
I wouldn’t get too hung up about Crystal Ball right now, though. It’s a great tool, and one that you can use to try to estimate how interacting uncertainties may accumulate. It’s a much better approach than the old “factor of safety” rules of thumb we’ve always had. But this is a first course only.

Comment: Believe it or not, the muddiest part of this module had to do with drawing the network (ADM) by hand. In other words, I had a hard time using “dummy” lines to represent certain activities (concurrent activities) used in certain examples throughout the text. For instance, looking at Figure 8-10 on page 381, how would I know which of those two diagrams is correct (since I haven’t seen the WBS)? If I took the first illustration and rotated it 90 degrees counter-clockwise, then the diagram would look correct (flowing from left to right and illustrating predecessors accordingly). I halfway understand the differences howbeit the subtleties of these differences seem insignificant in some circumstances.

Instructor reply: I am not the best person to explain this, but I can see what they’re doing here. Looking at the caption, the “RIGHT!!!” diagram makes explicit the precedence relationships that are assumed here. In the “WRONG!!!” diagram, a dummy leg goes from node 1 to node 2. A second dummy node goes from node 2 to node 3. The implication is that activity “a” as well as “b” and “c” precedes “f”; however, in this example, “f” is only supposed to depend on “b” and “c”. However, “a” precedes both “d” and “e” and “e” is preceded by both “a” and “b”; thus, the dummy leg must be inserted as shown in order for the logical relationship to be correct. Again, this comes straight off the caption. Note that there is a typo in the paragraph on p. 380 where it’s referencing this diagram. The description really gums it up to the point where it’s incomprehensible without looking at it closely.

Comment: Project scheduling appears to be a dynamic, rather than static process, in that conditions affecting completion have the tendency to vary, especially over the course of a multi-year project. Is there a rule of thumb or recommendation on how often to generate new or updated networks over a project of long duration?

Instructor reply: PMI comes down hard on the side of updating project schedules and baseline schedules. The baseline schedule is what the planned schedule originally was, or if an approved change is made, it is an updated baseline that reflects the results of that change. In case where there is a significant change to the project timeline for whatever reason, the schedule should be updated and a revised schedule (with a revised finish date) must be approved. Of course in many cases this slippage can be to the detriment of a PM’s career. The updated schedule will need to reflect the new network, if there is a change to the critical path. So, the answer is that there is no rule of thumb as to how often to do this, except the answer is “never” if (as is desired) the project schedule remains intact. But if something changes, the Gantt chart and network should be updated per PMI. Also of note is that it is often good practice to update the Gannt chart with the actual vs. the planned progress.

Of course, in large projects it’s common for the critical path to flip from one thing to another, and it can be really very taxing on everybody to monitor all this. See Dr. Perkins’ comments on just this problem in the Old Closure of Module 9.

Comment: Muddiest thing: Setting up predecessors. I believe it can get really complicated in a big project and take a long time. Already problem 11 caused some brainwork. Is there a way to do it graphically in MS Project? Just creating the nodes and then click with the mouse to draw arrows between them!?!

Instructor reply: You can use the “link tasks” button on the toolbar to graphically connect activities in the Gantt chart. You can’t do it on the network diagram to my knowledge. There are other programs (a simulation language called iThink comes to mind) where you can and do link “nodes” using a mouse as you’re suggesting. Maybe someday someone can create an aftermarket product that does what you’re suggesting. I think a lot of us would appreciate it.

Comment: I really don’t like AOA, I much prefer AON, it seems much more intuitive, for me.

Instructor reply: I think this is personal preference. The advantage of AON is that the software packages go in that direction. The appeal of AOA, for me at least, is that there is something that feels a little better about the use of an arrow length to represent a duration, even though you can’t draw the arrow to a time scale in any practical sense. Someday I plan to read an old text of mine that gets into arrow diagramming methods so I can get a little more insight as to why AOA (a/k/a ADM) might have advantages.