Projects in manufacturing can be of many shapes and sizes. They may involve the development of new products, the design or improvement of manufacturing facilities, the application of new technologies or standards, and the actual manufacturing of components for a specific program. Each of these applications requires planning and control. However, there is more than one approach toward such planning and control. Choosing the appropriate methodology and tools is crucial to success of the project itself.
The critical path method, of course, is the best known and most widely employed method for project planning and control. For non-production type projects, it continues to be an appropriate process for planning, scheduling, resource loading and budgeting the work, and for tracking and control. I would not hesitate to apply this process to most new product development and facilities type projects.
However, there are several manufacturing situations that would benefit from other methodologies and for which the CPM effort would surely flounder and die. Even when CPM and critical path software is appropriate, there are certain conditions in many manufacturing projects that would benefit from some of the special features available in only a few commercial CPM or allied products. Our first example describes one of these circumstances and a product that offers an unusual capability to support critical material resource supply and demand situations.
When your project involves actual production, your planning and control interest should stress productivity and critical flow, rather than tracking individual parts. Here too, we often find traditional critical path methods put into use—often to the detriment of practical planning and control. It is important to remain open to alternative approaches. The better methodology for this is a process developed over 50 years ago—Line of Balance. In our second example, we define the LOB process, and how it can be employed in typical manufacturing situations.
Defining and analyzing the critical flow through a manufacturing process is often a hit-and-miss proposition. The ability to evaluate and optimize the process has just become easier due to the availability of a new PC-based tool. In our final example, we describe this tool and how it supports simulation and analysis of various process options.
Producing and Consuming Materials Resources
Virtually all popular project management software products provide resource scheduling and leveling. The typical use of these functions consists of assigning resources to tasks, specifying the quantity of resources available, and rescheduling tasks so as not to exceed resource limits. Scheduling problems in the manufacturing arena tend to be more productivity-oriented (except in new product and prototype development situations).
An extension to the traditional resource scheduling and leveling functions, available in a few critical path scheduling products, is the ability to define consumable or depletable resources. Instead of defining a rate of resource availability (i.e., three steamfitters for eight hours per day), a consumable resource is defined as a quality that is available for depletion, as the resource is assigned to tasks.
A unique extension to this capability, available in SuperProject 3.0, is the ability to specify tasks that produce resources, as well as consume them. Consider, for example, a series of tasks that involve pouring concrete. You may specify the total amount of concrete required for each task, or the quantity required for each time period. Then you may also create a task representing the delivery of concrete to the job site, and specify the rate of delivery (such as cubic yards per hour). SuperProject 3.0 provides a predesigned view that shows the tasks at the top and the material use and inventory at the bottom. You can clearly see if the scheduled use exceeds the inventory. If it does, you can invoke the resource leveling function to reschedule the tasks to stay within the delivered quantities.
I applied this capability to a manufacturing situation, where the rate of fabrication of a scarce material was delaying the execution of a project. This application consisted of scheduling several projects involving the manufacture of fiber-optic cable and supporting materials for laying transoceanic cable systems. Before a ship can leave port to execute the cable laying portion, the specified amounts of cable and supporting materials had to be on board. In this case, the production of power cable was found to be the limiting item. Current capacity was 24 km per day. Additional production lines were scheduled to add 6 km per day on 12/2/91 and another 6 km per day on 2/3/92.
Using the special capabilities in SuperProject 3.0, we created three power cable production tasks. Produce Power Cable A, started on 10/4/91, had a resource assignment of 24 km per day—entered as a producer resource, rather than a consumable resource. Power Cable B, started on 12/2/91, had a resource assignment of 6 km per day—entered as a producer resource. And Power Cable C, started on 2/3/92, had a resource assignment of 6 km per day—entered as a producer resource (see Figure 1).
Next, we created four tasks (one for each project) representing the delivery of the required power cable to each ship. We assigned the resource “Power Cable” to each task, as a consumable resource, in the required quantities for each project. (We also created tasks to represent the other work associated with the projects.)
Figure 1. Resource Tracking in SuperProject 3.0
When we invoked the resource leveling function, SuperProject added power cable to inventory, based on the three producer tasks, until the amount available equaled the amount required for the first ship (in this case, 680 km). It then actually showed the date that the ship would have all materials on board and would be ready to sail. Inventory was then automatically reduced by the demand of Ship A (680 km) and continued to build based on the programmed production rate. With this capability, the manager could experiment with project priorities, vary production rates, substitute materials, etc., and immediately view the effect of such actions on the sailing dates, and on project completion.
While it is not too common to see critical path scheduling programs applied to production-oriented manufacturing situations, it should not be difficult to see the potential applications of this unusual capability.
Line of Balance
There have been two situations, one in the ’60s and one in the ’70s, where I had to put my job on the line to defend my position to apply the Line of Balance technique to a project. In both cases, management, not being aware of alternatives, had requested the application of traditional CPM techniques. I will describe the second incident here.
In 1974, the General Electric Company's Gas Turbine Division was awarded a contract to manufacture equipment for the USSR's expanding gas pipeline program. GE was to manufacture, and deliver to ships, equipment consisting of gas turbines, gas compressors, and all related hardware for 65 compressor stations. Equipment was to be manufactured by several vendors, in eight countries. The contract called for over 2,000 final equipment shipments, and the scheduling and monitoring of many thousands of key manufacturing events.
Even I, initially, considered using CPM because of its widespread familiarity and application as a program management tool. However, for this project, the act of following significant activities, each with a specific elapsed tine, for so many components, would have resulted in personnel spending an excessive portion of their time manipulating data to meet the exacting demands of the CPM process. There was really no need to know where each piece was at any time. Nor was it necessary to pre-associate each part with a particular end unit (as the CPM would have required). What we needed to know was how many units were required at any time to support a set of end objectives. Let's look at how the LOB method supported these needs, with just a small portion of the effort.
Figure 2. LOB Three-Step Operation
LOB is a management tool involving planning, scheduling, and control. It is a graphical method of scheduling that directs attention to critical activities and schedule non-conformance, and provides for management by exception.
The particular objective of LOB is “to develop a progress study of the significant operations of a project so as to be able to examine the actual progress at periodic intervals during the life of the project.” This entails the integration and monitoring of the flow of materials, components, sub-assemblies, etc., through manufacturing and assembly, in accordance with the requirements for delivery of end items. It employs a three-step process: identify the objectives; establish a plan to meet the objectives; measure progress against the plan. In LOB, this is accomplished by using a three-step graphic operation consisting of an objective chart, production plan, and progress chart (see Figure 2).
The objective chart plots the cumulative quantities of deliverable end items against a calendar. It is used with the production plan to establish required quantities for all interim operations.
The production plan is based on a logic diagramming technique similar to that of CPM. In LOB, it is typical for the production plan to be the same for all identical units, and for the logic diagram to be drawn against a time scale. The production plan provides a set of lead times for key events; that is, the total number of days or weeks that a task must be completed prior to the completion date of the end item. When the cumulative end quantities of the objective chart are applied against the production plan, we derive the number of units required to be complete, for any event in the production plan, for any time during the project. This provides the data for the line of balance that is drawn across the progress chart.
There are two parts to the progress chart; the line of balance bar and the progress columns. The required quantities at a specific point in time are shown graphically in the progress chart as a stepped horizontal line across the chart. This is the line of balance, and it identifies a position on the graph that respective actual quantities should reach or exceed at that measurement time. The actual quantities to date are shown graphically on the progress chart as heavy vertical columns.
The objective chart and the production plan are the planning operations of the LOB technique, and generally do not change during the program. The progress chart, however, depicts the requirements and status at a particular point in time, and must be updated each time that the schedule progress is to be examined.
For periodic analysis, you just compare the actual quantities for task completions to the planned or required quantities depicted by the line of balance. Columns that do not come up to the line of balance represent production deficiencies. Events having bars the farthest distance below the LOB are the critical operations. It is therefore possible to look at a single chart for a project—to find the critical operations—and to use this information to redirect efforts from non-deficient operations (bars above the LOB) to the delinquent operations.
This efficient LOB technique, while based on traditional logic diagramming and management by exception protocols, frees personnel from having to acquire date and time cycle data for each step for each unit, and from having to assign each activity to a specific unit. Implementing this technique on the GE Russian Pipeline project improved our communication with management and contributors, and reduced the time and effort required to process and monitor project progress. We were able to easily find the areas where production expediting was required to fulfill shipping deadlines. This was very important because the contract called for payments only when an entire shipload of components was on board and the ship was able to sail.
Although LOB is primarily a graphic technique, the computer can be effectively employed to compute the line of balance for each period. We wrote our own code for the GE application, using GE's time sharing computer system. The computerization process is very simple (way, way less complex than today's CPM programs). Yet, I do not know of any commercially available LOB programs on the market today. (TimePhaser —not currently supported—had a modified LOB capability, built into a traditional CPM process.)
Process Simulation and Flow Analysis
Speaking of the lack of computer solutions for management processes, it's even harder to believe that “process flow analysis”—something that every manufacturing firm (and most others) do on a regular basis—has been ignored by the software industry. That is, until now. Scitor, producer of Project Scheduler 6 (critical path software), has released a breakthrough product called Process Charter – “The Flow Charter with Brains.” It is the first time that any process analysis software has been available at a popular price ($595).
It is predicted that the Windows-based, easy-to-use, Process Charter will do for the Process Flow analysis application what VisiCalc and Lotus 1-2-3 did for the spreadsheet application. Process simulation and flow analysis is a natural for the computer, and the effort by Scitor to bring this application to the PC is receiving universal recognition from both the developer and user communities.
Here are the typical functions that are employed in doing a process flow and simulation:
Figure 3. Process Analysis with Scitor's Process Charter
1. We start with mapping the process. That is, we analyze the process and create a visual picture or flow chart of the process. For this step, Process Charter provides a capability similar to ABC Flow Charter or Visio, wherein we create figures depicting flow steps or decision points and the path of the flow. When branching from any node, we can specify the probabilities of going down each path. This can account for rejection rates, machine downtime, and other process options. By varying these values, we can analyze the effect of varied quality specifications or process options.
2. Next, we define resources: people, materials, machines. For each resource, we can specify resource name, type of resource, type of use (reusable or consumable), quantity, normal cost rate, normal work period, O/T rate, and max O/T time.
3. Now we can assign resources to activities or elements of the flow chart, and specify several properties for the activity and assignment. These options include choosing calendars, priority levels, maximum queue size, time options (what to do at the end of an active period as defined by the activity's calendar—such as suspend, finish task, never extend, or finish all inputs), and several other parameters. Calendar specifications define the available working period for an activity. This can be a five-day week, every Friday, every last day of the month, etc. The activity spans for each workday can also be defined.
4. Step four is simulation of the process. When we select “Process Run,” we actually execute the process defined in the model. If we wish, we can observe the simulation. Process Charter will display different colors at the nodes as they are activated during the run, with each color indicating conditions such as queuing and bottlenecks. If we wish, we can even step through the process one cycle at a time. But this is not necessary, because an analysis report is produced at the end of the run.
5. The summary tab will provide cost breakdowns, time breakdowns, resource blocks and delays, etc. We may also specify a set of “key values” to produce a defined spreadsheet of analysis data.
6. The key values spreadsheet can hold the results of several runs so that various alternatives can be compared.
Following these steps, it is possible to model virtually any process and simulate and analyze the process. Trying various iterations of the model will lead to defining the best alternatives to meet various time, cost, and quality objectives in most manufacturing situations.
Here's one example of the use of this technique and tool, in a manufacturing situation:
The Company: ACE Computer Company manufactures and assembles microcomputers. ACE's Assembly Division processes orders for various configurations and ships completed orders.
The Analysis Problem: ACE Computer offers two CPU models. They get orders for either their base system or their fast system. Either of these systems can be orderd with a multimedia option. Their experience indicates that only half the basic systems go multimedia, where 75 percent of the fast systems have the multimedia option.
Figure 4. Resouce Allocation with Process Charter
During assembly, technicians are assigned to the hard stuff and ordinary assemblers handle the basic assembly.
There is a 1-in-20 chance that a completed system will not check out, and the Tech will have to diagnose the problem and repair it.
Systems gather up at shipping, where once a day they are picked up.
Parts get used up during the process. They are refreshed once a day at 10:00 when a shipment arrives from the warehouse.
The Process Charter Model and Solution: Figure 3 shows each of the steps in the process, from receipt of order to shipment. In most instances, there are two sets of flow lines, representing basic and fast systems. On Steps 3, 6, and 7, we placed a 50/50 probability, again representing basic and fast systems. Out of Step 8, we define a 50 percent probability for multimedia components, for the basic systems, and a 75 percent probability for multimedia components, for the fast systems. On Step 11, we send 5 percent of the systems off for rework.
When we process a “run,” the orders are split along probability lines, and components are drawn from stores as defined in the resource assignment. Technicians and assemblers are assigned to activities for specified times. The run simulates the actual factory process, evaluating times and resource requirements. Spreadsheet-type reports indicate average times at each step and how often each step is critical to the process (bottleneck).
If, for instance, the simulation shows that the assembly process is bogging down at the multimedia step, we can evaluate several alternatives. We might consider increasing labor resources or decreasing the time (by improving the process). Or we might even consider raising the cost of the multimedia option and changing the probability formula. Using the same basic model and changing the resource, time, or probability values for the multimedia step, we can evaluate the effect of these alternatives. The diagnostics will show time, effort and costs. Figure 4 shows the resource allocation, in percentage, for each element of the flow chart. Obviously, the options for evaluation are virtually limitless.
Does It Work? One of the early users of Process Charter is Modern Hard Chrome, in Warren, Michigan. This manufacturer of chrome coatings is using Process Charter to simulate and analyze their chrome-plating process. After modeling the process in Process Charter, Charles Nicholl of Modern Hard Chrome found that he could improve the efficiency of the process by expanding the capacity of a single piece of machinery in one facet of the process. Mr. Nicholl noted that using this tool to model and analyze the plating process gave him a much better understanding of just what was happening during the process and of what effect the key variables had on that process.
Summary
Virtually every type of manufacturing project can benefit from the application of structured planning techniques and tools. The effective, modern manufacturing manager or planner will maintain an up-to-date toolbox, consisting of many varied tools, and the knowledge of such tools and techniques so as to be able to select and apply the best tool for the job. Try to remember not to fall into the common trap of thinking that traditional CPM is the only technique available for project planning and control. Hopefully, I have proved that there are viable alternatives that are both more effective and more efficient for selected situations.
For more information on SuperProject, and Process Charter, contact the vendors. A good reference for information about Line of Balance is the Scheduling Handbook by James J. O'Brien, McGraw-Hill, 1969. Details on the LOB application discussed in this article, were published in the proceedings of the 5th Internet World Congress, Birmingham, U.K., 9/16/76 (Levine, Aliberti, and Ford). Don't let the ancient dates mislead you. These are still reliable references.
Harvey A. Levine, principal, The Project Knowledge Group, Saratoga Springs, New York, has been a practitioner of project management for over 30 years and is a past chairman of the Project Management Institute.