The Decision Sciences Journal of Innovative Education

Introducing Project Management Concepts using a Jewelry Store Robbery

By

Edward D. Walker II, Ph.D.

Department of Information Systems and Logistics

College of Business Administration

Georgia Southern University

P. O. Box 8152

Statesboro, GA 30460

(912) 681-5085
Introducing Project Management Concepts using a Jewelry Store Robbery

Introduction

This paper describes the use of simple hands-on case in undergraduate classes to teach the concepts of a traditional PERT/CPM and Critical Chain project management techniques.  A simple six activity project is used to illustrate: the identification of project activities, development of activity time estimates; the development of the project network; the development of the PERT/CPM critical path and the probability of on-time completion using the critical path; the development of the critical chain and the probability of on-time completion using the critical chain.  The case can be readily extended to discuss the effects of activity crashing, as well as, the proper placement of buffers within the critical chain network.  The overall objective is simply to illustrate the development and use of the critical chain in single project management.

Identifying the Project Activities

In this exercise I ask the students to assume that they are interested in “knocking over” a jewelry store.  [Some instructors may balk at the idea of using a robbery as an example.  The concepts can be taught using a different story line for the project network—one possible project would be writing a group paper for class.]  The students are divided into groups of three—representing the three would-be felons. They are told that they are to plan a cat burglary rather than simply walking into the business with weapons drawn.  They should plan to do this at night because the police have a longer response time and because the night patrolman comes by every 50 minutes.  The store has an external alarm; a jewelry safe; an office alarm; and a safe full of securities in the office.  They are given the diagram in figure 1 and asked to develop a list of all of the activities that need to be completed in order to completely clean out the store.  After they are given a few minutes to develop the list of activities, compare their lists to the following: A, disarm the exterior alarm system; B, disarm the office alarm system; C, crack & clean out the office safe; D, crack the jewelry safe; E, clean out the jewelry safe; and, F, pick up the loot & exit.

Insert Figure 1 about here

Developing the Activity Time Estimates, Identifying Activity Precedence Relationships and Developing The Project Network

After the list of activities has been developed, I ask whether the job can be completed successfully.  Immediately the students recognize that they must have some estimate of the activity durations.  I provide the following activity durations: 20 minutes to disable the exterior alarm system; 7 minutes to disable the office alarm system; 7 minutes to crack and clean out the office safe; 14 minutes to crack the jewelry safe; 4 minutes to stuff the contents of the jewelry safe into bags; and 10 minutes to gather all of the loot and exit the premises.  

The total of all activity time estimates is 62 minutes.  Ask the students if they should abort the plan since the total time of the activities (62 minutes) is greater then the allowable time (50 minutes).  Students quickly recognize that all activities do not need to be performed sequentially, but rather, that that some activities can be accomplished in parallel whereas others must be in a specific technological order.  The next step would be to decide the proper order in which to proceed.  The exterior alarm must be disabled before any of the other activities can begin.  But once inside the store, they realize that the office alarm and safe can be attacked while the jewelry safe is being opened.  I tell the students that they are all in this thing together, and stipulate that everyone must exit the building as a group—either they all are caught or they all get away clean.  After some discussion about how to create an ordered list of activities, I introduce the activity-on-node method of diagramming a project.  (See figure 2.)

Insert Figure 2 about here

Determining The Critical Path And The Probability of On-Time Completion

There are two possible paths through this network: the upper path A-B-C-F and the lower path A-D-E-F.  These paths will take a total time of 44 and 48 minutes respectively.  Each path is well within the allowable time.  I then show them how to calculate the slack associated with each activity—determining the early start/early finish time and late start/late finish time, respectively, for each activity.  As each activity on the lower path has an associated slack value of 0, the lower path is the PERT/CPM critical path.  They usually decide that since the patrolman makes rounds every 50 minutes they have ample time in which to accomplish the project. 

Since the penalty for being late on the project is substantial (i.e. going to prison), they must examine the probability of on-time project completion.  For this the students will need optimistic (O), pessimistic (P), and most likely (ML) time estimates for activity duration.  [We cannot know until the project is completed how much time the activities will actually consume; however, when asked how long an activity might take, most people can, by drawing on prior experience, provide best-case, worst-case, and most-likely estimates.  I try to avoid getting into the technical differences between PERT and CPM when introducing project management.] PERT/CPM activities are usually assumed to follow a Beta distribution.  These data are in Table 1.  Given the calculations, the students often decide that the risk is worth the reward and that a 7.35% chance of being caught is acceptable.

Insert Table 1 about here

Identifying the Critical Chain and the Probability of On-Time Completion

In this case, the project is to be performed by three persons selected for their special abilities.  One knows about and can defeat any alarm system.  The second can crack open any safe, and the third can carry the great weight of the loot.  There are three resources available for use.  Ask the students to assign resources to each activity, and they realize that the safe-cracking accomplice is needed in two places simultaneously. 

Now that we recognize that capacity must be considered when developing a project network, we then compute the effects of this allocation.  Activities A and B are assigned to the alarm specialist; activities C and D are assigned to the safe cracker; and, activities E and F are assigned to the “pack mule.” One way to complete the project is to complete activity D before activity C.  I use a dashed arrow to indicate the resource precedence relationships as opposed to the solid arrows used to indicate technological precedence relationships (draw a dashed arrow from D to C in figure 2.)

Using this diagramming method, it is simply a matter of applying the same logic used to calculate the PERT/CPM critical path to calculate the critical chain.  However, we must consider not only the technological relationships within the network but also the resource relationships.  We find then that the critical chain is A-D-C-F.  This chain has an expected total duration of 51 minutes.  Using the same optimistic, pessimistic and most likely estimates of activity duration just in table 1, we find that the probability of on-time project completion drops to 25.78% (Z = -0.65).  [Instructors can discuss the importance of resource allocation here.]

Activity Crashing and Buffering the Network

The concept of activity crashing can be discussed at this point.  We know that the project will most likely fail if attempted as currently planned.  I provide the students with various crashing options for activities A, B, C, and D in the form of the purchase of electronic equipment to defeat the alarms or open the safes.  By asking leading questions, I get the students to realize that crashing activity B is not helpful as it is not on the critical chain and that we should choose the least expensive option or options with respect to cost-per-minute-crashed.

Once the project has been crashed to the point that the probability of on-time completion is acceptable, we discuss the necessity of buffering the project. By answering leading questions yet again, the students come to understand that we have already inserted a project completion buffer—the time between the due date and the expected completion of the project—by increasing the probability of on-time completion to something greater than 50% and that the non-critical activities B and E should be started sometime before their late start time to allow for pessimistic activity completion—the insertion of feeding buffers.

Summary

This simple case has been used effectively to teach abstract and complex concepts of project planning and control to undergraduate.  Students enjoy playing the game and trying to figure out how to design and manage the project network.  Most students are capable of creating a critical chain project network without any or much intervention by the instructor.  Evidence of the effectiveness of this case is provided in table 2.

Insert Table 2 about here
Table 1: PERT/CPM calculation of the probability of on-time project completion.

Critical path (A-D-E-F) = 48 minutes

Allowable time = 50 minutes

Variance on critical path = 1+.45+0+.45 = 1.90

Standard deviation of the critical path = = 1.378

Z = (50-48)/1.378 = 1.45

92.65% chance of completion on or before 50 minutes.

Table 2: Survey results of two undergraduate classes.