B. HUSCHB. HUSCH, a former member of FAO's Forestry Department, is now in the Economic Analysis Division. He is the author of several publications on forest mensuration and inventory
WITHIN RECENT YEARS there has been a series of important developments in management science which are gradually being employed in forestry with impressive results. These developments which include such esoteric-sounding techniques as systems analysis, mathematical modelling, mathematical programming (linear, nonlinear and dynamic) and simulation have been used in a scientific approach to decision-making often called " operations research " or " operations analysis. "
Within the field of operations research, one particular group of techniques, which can best be called network analysis, has been developed and has proved to be a powerful tool for project formulation and execution. It is only a little more than a decade ago that the first form of network analysis was developed in the United States and given the name Program Evaluation and Review Technique (PERT). Since then a number of similar or related methods have been developed, among which the best known is the Critical Path Method (CPM). However, the basis for all the variations is the graphic model of a project called a network, and consequently the collection of methodologies is referred to as network analysis.
Following its earliest application for weapons systems development, network analysis. was soon applied to construction projects, with dramatic reduction in expenditures of time and money. It is only recently, however, that this management tool has been applied to forestry projects (Davis, 1968; Husch, 1969).
FAO, in its international assistance projects in agriculture, forestry and fisheries has begun to apply the approach to project formulation and execution. Although the following discussion shows the application of network analysis to forestry development projects, the method is applicable to any problem or project involving a number of interrelated and time-dependent activities.
To be properly applied and to yield best results, network analysis must be considered as only part of a more inclusive project-management process. Thus, before proceeding to the intricacies of a graphic model it would be worthwhile to examine the management process and see where network analysis fits in.
The fundamental stages in the management process for formulating and executing a project are summarized diagrammatically in Figure 1. This simplified diagram indicates that the first and most important requirement in preparing any project is the determination and definition of the objectives one hopes to achieve; subsequent accomplishments should be measured and evaluated in comparison to them.
FIGURE 1. - The process of formulating and managing a project.
The process then moves to the planning stage, which includes the determination, sequence and interrelationship of the activities required to achieve the objective of the project. It is in this stage that the network diagram is used which gives network analysis its name.
Figure 1 shows that the process then moves to the third stage of scheduling in time units. The goal of scheduling is to produce a timetable of the activities of the project which will be realistic for the resources available and their desired manner of employment, and consistent with the desired completion date for the project.
Project Operations then get under way, as shown in the fourth stage of Figure 1. During operations there must be a continuous check of actual operations against those planned; this allows for early detection of potential project difficulties and permits the seeking of solutions prior to serious consequences.
As project operations continue, changes in the plan may become necessary. The evaluation of the effects of these changes constitutes the fifth step in the process. The changes contemplated may involve scheduling the planned tasks of the project, or even the project objectives. The suggested changes are evaluated and a decision is made on the action to be taken to modify the project. This cyclic action is indicated in Figure 1 by the return arrows. Feedback may return to any or all of these three stages depending on the changes required. The process is dynamic and continuous, recycling as many times as needed during the project and ceasing only at project completion.
What is the role of network analysis in this management process? It is a procedure which utilizes a rigorous, logical approach to project formulation and execution, stressing advance planning and the identification of critical elements.
Network analysis in the management process can be summarized as consisting of the following steps:
(a) Specification of detailed project objectives.(b) Selection of project activities to achieve each objective and estimates of time and resources required for each activity.
(c) Development of a graphic model, in the form of a network diagram, indicating the planned sequence, interdependencies and interrelationships of the project activities.
(d) Calculations of time estimates for activity and project completion and the identification of critical paths in the network.
(e) Study and evaluation of the project model (i.e., the network) to see if it logically ties together all parts of the project and to see if objectives can be reached within acceptable time limits and use of resources. From the evaluations and subsequent modifications a schedule of project activities and events is then prepared.
(f) Use of the network and schedules to monitor project progress, comparing expected results, time and costs with actual ones, identifying problems and taking corrective action or making modifications.
Too often network analysis has been considered to he simply a means of organizing project activities, concentrating on the drawing of a network diagram and time calculations. To realize the full usefulness of network analysis one must start with a searching consideration of project objectives. Only after objectives have been decided upon and described and. the work necessary to achieve them specified should we proceed to the modelling phase of preparing a network diagram.
A useful procedure for the analysis and selection of project objectives, called " objective structuring," considers that the objectives of a project form several levels of explicitness in the form of a pyramid. At the first level or apex, the prime objective states the desired result of the project in broad terms. This single objective can then be subdivided into a series of supporting objectives at increasingly specific levels. The achievement of the combinations of supporting objectives at each level means the achievement of the project objectives at the next higher level; for example, the prime objective may be subsivided into several more detailed specific objectives at a second level and each of these may be then further subdivided into a third level of even more specific and limited objectives. The number of objective levels should be carried to the extent necessary for the clear identification of the activities or tasks needed to accomplish them. This structuring of project objectives is illustrated by Figure 2, prepared for a hypothetical forestry development project in an unnamed country. Figure 2 shows graphically the order of increasingly specific objectives as one descends to lower levels; the accompanying key defines the objectives. In practice, all objectives at each level are studied for possible subdivision, descending no further than is necessary to establish an end which can be met by no more than a few clearly defined tasks on the project. These, called the working-level objectives, are indicated in italics in the key.
The structure is not a flow chart but rather an organization of related but distinct objectives. The objective structure is not concerned with arriving at these goals but is simply a method of identifying what must be achieved. The strategy and work necessary to attain these objectives follow during the planning and scheduling steps of the management process.
FIGURE 2. - Example of the structuring of project objectives.
(the working-level objectives are shown in italics)
1. Assistance to the government in the preparation of a comprehensive forestry and forest industries development plan2. Information on the country's forest resources (species, sizes, location, accessibility and value)
21. Reconnaissance information on native forest resources
22. Accumulation and analyses of existing forest resource data221. Accumulation and analyses of existing inventory information on plantations
222. Accumulation and analyses of existing inventory information on native forests23. Detailed inventory information on pilot management area
24. Detailed inventory information on selected forest areas of high potential
3. Information on consumption and marketing of wood products
31. Knowledge of past wood consumption and forecast of future trends
32. Knowledge of exports and imports of wood products and predictions for the future4. Determination of the role of forest plantations
41. Determination of sites for plantation establishment
42. Determination of suitable tree species for plantations
43. Demonstration of plantation establishment
44. Knowledge of economics of plantation establishment, management and harvesting
45. Determination of techniques for plantation establishment5. Determination of measures to manage the native forest to obtain greater yields without adverse effects on soils
51. Knowledge of suitable forest management techniques
52. Preparation of management plans for native forests
53. Demonstration of forest management of native forests
54. Knowledge of agricultural practices suitable for combination with forest land use (e.g., cocoa)
55. Information on and recommendations for harvesting methods suitable in the native forests6. Provision of trained local personnel
61. Provision of on-the-job training
62. Provision of courses for project personnel
63. Establishment of fellowship programme7. Assistance to the government in the preparation of appropriate legislation
71. Study of existing rural legislation and recommendations
72. Study of present institutions and recommendations8. Assistance in the preparation of a forest industries development plan
TABLE. - EXAMPLE OF ACTIVITY LIST TO SATISFY WORKING-LEVEL OBJECTIVES IN PLANNING A PROJECT (See Figure 2 for project objectives structure)
|
Working-level objective |
Activity |
|
Accumulation and analysis of existing information on native forests |
Assembling and study of existing forest resource data on native forests |
Accumulation and analysis of existing information on plantations |
Assembling and study of existing forest resource data on plantations |
Reconnaissance-type inventory information on native forest resources of country |
Reconnaissance of forests in the project area and
preparing of specifications for inventory to be carried out on a contractual
basis |
Detailed inventory information on selected forest areas of high potential |
Intensive inventory by contract of selected forest areas |
Detailed inventory information on pilot management area |
Intensive inventory by contract of pilot management area |
|
Knowledge of past wood consumption and forecast of future trends |
Wood consumption studies and forecasts |
Knowledge of quantities of past exports and imports of wood products and predictions for the future |
Trade and marketing studies of wood products |
Knowledge of economies of plantation establishment, management, and harvesting |
Plantation economics studies |
Determination of sites for plantation establishment |
Soil and ecological reconnaissance of the project area |
Determination of suitable tree species for plantations |
Establishing and operating nursery |
Determination of techniques for plantation establishment |
Demonstrating plantation establishment |
Knowledge of suitable forest management techniques |
Forest management studies and trials |
Management plans for native forest |
Preparing forest management plans |
Demonstration of management of the native forest |
Selecting pilot management area |
On-the-job training |
Local personnel participate in all activities |
Fellowship programme |
Granting and execution of fellowships for training and study tours |
Courses for project personnel |
Organizing project courses in forest inventory, plantation establishment methods, and forest management |
Report on rural legislation and recommendations |
Study of rural legislation related to forestry and forest industries |
Report on institutions and recommendations |
Study of institutions related to forestry and forest industries |
Forest industries development plan |
Preparing forest industries development plan |
|
Project Administration Availability of equipment and supplies |
Purchase of: vehicles |
Provision of international and local staff |
Recruiting of international and local staff (determination of staff requirements to follow the network analysis) |
Contracts for studies and provision of services |
Arranging contract for forest inventory |
Planning consists of preparing a scheme of action or procedure prior to operations, describing the work necessary to achieve the previously established objectives. It consists of two phases:
(a) A description of the tasks to be executed, the logical order or sequence in which they should occur, recognition of the necessary prerequisites before they can begin, and their interrelationships and interdependences with other tasks on the project;(b) Estimates of the duration of each project task, with its starting and finishing times, and the project completion time.
Network analysis provides a powerful tool for these two planning tasks. The first requisite of project planning using network analysis is to describe all the tasks which the project must complete to achieve the predetermined objectives. In network analysis terms, these tasks are called " activities. "
This list of activities must be carefully thought out. An example of such a detailed list for the hypothetical forestry development project, shown in Table 1, was prepared by first extracting all the working level objectives from the project objective structure of Figure 2 and listing them on the left side of the table. Under the " activity " side of the table are then shown all the activities considered necessary to achieve each objective. For more detailed planning, the list may be expanded and many of the activities divided into several subactivities. In addition, numerous other tasks of an administrative character required to permit executing this list of technical activities could be shown in greater detail. These may include the acquisition of equipment, instruments, aerial photographs and supplies, recruitment of staff, and so on.
Once the list of activities has been prepared, interrelationships and sequencing can be developed by means of network diagramming.
A network is a graphic model of the system representing the activities of a project. It is a logic model illustrating the relationships between the component activities and indicates the flow of operations from beginning to end of the project. The network for the hypothetical forestry development project, shown in Figure 3, consists of a series of arrows representing the activities that require time, manpower and other resources for completion. A description of the activities represented by each arrow is given in the first column of Table 2. A major feature of the network is its ability to show interrelationships and sequences, since it requires that one consider where activities can start and finish in relation to each other.
Each activity starts and finishes at a node which is called an event (the numbered circles). Each event represents an instant in time as either the start or completion of one or more activities.
The drawing of the diagram is an exercise in the kind of logical thinking that is of fundamental importance to project planning and operation. In simplest terms, the network diagram is begun by first deciding which activities can be started without completing any prerequisite activities. They are shown as arrows leading from the first event or initiation of the project. The end of each of these activities is also an event. The activities which can logically follow or are dependent on the first activities are then indicated as subsequent arrows leading from the previous events. This process is then repeated until all the activities have been incorporated into the network as arrows in their logical sequences and showing all dependencies. One should not expect the first effort at preparing a project network to produce an optimum or even satisfactory project plan; repeated revisions during the course of diagramming and even after completion will be required before a satisfactory network results. It is precisely in this scrutiny and revision that the value of the network diagram will become apparent. The procedure forces one to examine the logical and sequential relationship between all activities of the project; poor logic and impossible sequences become vividly apparent.
To understand the network diagram shown in Figure 3, a few simple rules should be followed:
1. Each activity is represented by only one arrow bounded at each end by an event represented by a small numbered circle. The event at the tail of the arrow represents the beginning of the activity and the event at the head indicates its completion.2. Length, curvature or direction of arrows have no significance.
3. Several activities can begin at a single event (a burst event). Similarly several activities can terminate at an event (a merge event).
4. Before any activity can begin all preceding activities must be completed.
5. Two kinds of arrows are shown in the diagram, one having a solid line and one a broken line. The solid line indicates an activity that requires time or resources to complete. The broken line is a " dummy " activity which uses no time or resources, but indicates a dependency relationship between events.
6. The length of the arrows is not scaled to the duration of the activity.
7. Events are numbered for identification purposes. An activity can be identified by citing the numbers of the two events bounding it, as shown in the second column of Table 2.
The network shown in Figure 3 is only a logic diagram and does not show any time calculations other than numbers above the arrows representing estimates of activity duration. A careful study will illustrate all the logical relationships which may be encountered, including dummy activities, breaking up of activities into phases, and merge and burst events.
ESTIMATES OF REQUIRED ACTIVITY DURATION
The network of Figure 3 is concerned only with sequences and interrelationships. This is only one part of network analysis in project planning; the other concerns time calculations of the start and finish times for activities and project completion.
Network analysis requires estimates of the time needed for the completion of each activity composing a project and designated by the arrows in a network diagram. These are shown in the third column of Table 2 and written above the activity arrows in Figure 3. All subsequent time estimates then utilize these duration times.
In the network analysis procedure used on FAO projects, a single best estimate based on experience and knowledge is used. The probabilistic approach of some network analysis procedures such as PERT, requiring three time estimates for each activity (i.e., most pessimistic, most optimistic and most likely), has not been found useful. In the course of subsequent project revisions, activity durations may be changed where required but always recognizing that, in turn, modifications of project objectives or resources to complete the activity may be required.
CRITICAL PATH
Viewing a project as a network makes it evident that there are numerous sequences of activities from beginning to end which must be completed to achieve all the objectives.
Some of these sequences may be independent of the rest of the activities in the project and others are interdependent. The same activity may thus form a part of more than one sequence. The combination of activities constituting a sequence from the first to the last event is called a path. The completion of the project depicted by a network requires the completion of every path, which is equivalent to the completion of every activity. The total time for the completion of a path equals the sum of all activities on that path. The path requiring the longest time is called the critical path and is made up of the critical activities. All other paths in the network are shorter. The length of time to complete any other series of activities will be less, but the earlier completion of these other sequences will not speed up total project completion. The difference between the length of time for the critical path and any other path is called the slack. Any delay in any activity on the critical path will result in a delay in the total project time. In contrast delays up to slack time can occur on noncritical paths without causing total project delay.
CALCULATION OF ACTIVITY AND EVENT TIME ESTIMATES
The most important time calculations in a network analysis are shown below.
For each activity:
earliest possible starting time;
earliest possible finish time;
latest allowable starting time;
latest allowable finish time;
total slack.
For each event:
earliest possible occurrence time;
latest allowable occurrence time.
The required calculations utilize only simple arithmetic. They can be carried out systematically in tabular form by hand or by computer. For simple networks of the type shown in Figure 3, a useful procedure is to prepare a modified diagram (as shown in Figure 4), which enables all the calculations to be carried out directly on the network.
Although the diagram and the calculations appear complicated they are basically very simple. The meaning of the various symbols and figures will become clear by studying the very simple project illustrated by Figure 5. First assume that there are no numbers on this diagram except for the identifying event numbers in the upper part of each circle and the duration time written above each of the arrows. All the remaining time estimates to fill in the empty spaces are calculated in two steps: a forward and a backward pass. For the forward pass a zero is entered in the earliest occurrence time for event number 1 (lower left side of circle). This earliest occurrence time is equivalent to the earliest start time for all activities beginning at this event. The time of duration of each activity is then added to the earliest start time to give the earliest finish time shown in the head box of each activity arrow. Where only one earliest finish time leads into one event, this time is equivalent to the earliest possible occurrence time for that event and is written in the lower left part of the circle. If more than one activity terminates at an event the largest of the earliest finish times for the activities becomes the earliest possible occurrence time for the event. All the succeeding calculations for the forward pass are then identical. Following this procedure one arrives at event 9 and the maximum of the earliest finish times of the activities terminating, 20, is entered as the earliest occurrence time for the completion of the project. The calculations then proceed to the backward pass. Using the assumption that the latest allowable time for the last event is the same as the earliest possible occurrence time, the value of 20 is entered on the lower right side of the circle. This time then also becomes the latest allowable finish time for all preceding activities. The duration time of each of the preceding activities is subtracted from this latest finish time to give the latest start time which is then written in the tail box of each activity arrow. When only one activity emerges from the preceding event, the latest start time for an activity is also the latest allowable occurrence time for the preceding event and the same value is written in the lower right side of the circle. If more than one activity emerges from a preceding event, the minimum of the latest start times is the latest allowable occurrence time and is transferred to the lower right side of the circle. The value indicated for the latest allowable occurrence time in the circle then becomes the latest finish time for the next preceding activity(ies) and the calculations continue backward until the latest allowable occurrence time for the first event is determined. If all arithmetic is correct, this should be a zero.
TABLE 2. - EXAMPLE. OF AN ACTIVITY LIST FOR PROJECT ANALYSIS
|
Activity |
Identification number |
Estimated duration (months) |
Prerequisite activities |
|
Establishing project headquarters |
10;20 |
1 |
- |
|
Acquiring equipment: |
|
|
|
|
1st phase - vehicles and soil survey |
10;30 |
3 |
- |
|
2nd phase - inventory and nursery |
30;50 |
5 |
(10;30) (20;30) |
|
3rd phase - planting |
50;220 |
7 |
(30;50) |
|
4th phase management, logging and transport |
220;240 |
3 |
(50;220) |
|
Assembling and study of existing forest resource data on plantations |
20;40 |
2 |
(10;20) |
|
Assembling and study of existing forest resource data on native forests |
20;60 |
3 |
(10;20) (10;30) |
|
Preparing programme of project courses |
20;70 |
1 |
(10;20) |
|
Forest inventory training course |
70;80 |
2 |
(20;70) |
|
Plantation establishment training course |
80;120 |
2 |
(70; 80) |
|
Forest management training course |
120;130 |
2 |
(80; 120) |
|
Soil and ecological reconnaissance of the project area: |
|
|
|
|
1st phase |
30;40 |
3 |
(10;20) (10;30) |
|
2nd phase |
40;430 |
9 |
(30;40) (20;40) |
|
Reconnaissance of forests in the project area and preparation of specifications for inventory to be carried out on a contractual basis |
60;90 |
1 |
(20;60) (10;30) |
|
Negotiation of inventory contract for both reconnaissance and intensive inventories |
90;100 |
3 |
(60; 90) |
|
Reconnaissance inventory of native forests of project area |
100;110 |
6 |
(90;100) (30;50) (70;80) |
|
Selection of plantation area |
40;190 |
1 |
(20;40) (30;40) |
|
Establishing field headquarters and nursery |
190;210 |
2 |
(80;120) (30;50) (40;190) |
|
Nursery operations: |
|
|
|
|
1st phase |
210;230 |
2 |
(190,210) |
|
2nd phase |
230;380 |
8 |
(50;220) (210;230) (190;200) |
|
3rd phase |
380;410 |
6 |
(230;380) |
|
4th phase |
410;480 |
5 |
(380;410) |
|
Select trial planting sites |
190;200 |
1 |
(40;90) (30;50) |
|
Species trials: |
|
|
|
|
1st phase |
230;390 |
8 |
(50;220) (210;230) (190;200) |
|
2nd phase |
390;400 |
6 |
(230;390) |
|
3rd phase |
400;470 |
4 |
(390;400) |
|
Tests of planting methods and equipment |
230;420 |
8 |
(50;220) (210;230) (190;200) |
|
Demonstrating plantation establishment: |
|
|
|
|
1st phase |
420;430 |
6 |
(230;260) (230;420) |
|
|
|
(230;390) (230;380) |
|
|
2nd phase |
430;450 |
5 |
(420;430) (290;300) |
|
|
|
(290;310) (290;320) |
|
|
|
|
(40;430) (380;410) (390;400) |
|
|
Preparing planting programme |
430 440 |
1 |
(290,320) (290,310) |
|
|
|
(290;300) (420;430) |
|
|
|
|
(40;430) (380;410) (390;400) |
|
|
Adjustment of contract for intensive inventory of selected forest areas and pilot management area |
110;150 |
1 |
(100;110) |
|
Intensive inventory by contract of selected forest areas |
150;270 |
10 |
(110;150) |
|
Intensive inventory by contract of pilot management area |
160;250 |
2 |
(110;150) (110;160) |
|
Wood consumption studies and forecasts |
170;250 |
4 |
(10;20) |
|
Negotiation of contract for wood testing |
110;140 |
3 |
(100;110) |
|
Wood testing by contract |
140;270 |
8 |
(110;140) |
|
Trade and marketing studies of wood products |
170;180 |
4 |
(10;20) |
|
Plantation economic studies |
230;260 |
4 |
(50;220) (210;230) (190;200) |
|
Study of rural legislation related to forestry and forest industries |
290;310 |
3 |
(10;20) |
|
Study of institutions related to forestry and forest industries |
290;320 |
3 |
(10;20) |
|
Selecting pilot management area |
110;160 |
1 |
(100;110) |
|
Forest management demonstration |
270;370 |
9 |
(140;270) (150;270) (250;270) |
|
Forest management studies and trials: |
|
|
|
|
1st phase |
250;270 |
6 |
(120;130) (160;250) |
|
|
|
(170;250) (170;180) (220;240) |
|
|
2nd phase |
270;340 |
4 |
(140;270) (150;270) (250;270) |
|
3rd phase |
340;360 |
4 (270;340) |
(270;340) |
|
Studies of agriculture forestry relations |
290;300 |
6 |
(10;20) |
|
Preparing forest management plans |
330;440 |
5 |
(250;280) (140;270) |
|
|
|
(150;270) (250;270) |
|
|
|
|
(290;320) (290;310) (290;300) |
|
|
Logging and transportation studies |
250;280 |
6 |
(120,130) (160,250) |
|
|
|
(170;250) (170;180) (220 240) |
|
|
Logging and transportation demonstration |
280;350 |
10 |
(250;280) |
|
Preparing forest industries development plan |
440;490 |
6 |
(270;340) (330;440) (430;440) |
FIGURE: 3. - Example of Network Diagram for Project - "Development of Forest Resources."
FIGURE: 4. - Network Analysis, UNDP Project - "Development of Forest Resources."
FIGURE: 5. - Direct time calculations on the network diagram.
The only figures now missing are those for the small circles in the middle of each activity arrow. This space is for total slack and the values can be determined concurrently with the backward pass calculations. There are several ways of determining the value of total slack, but the simplest and most easily interpreted is that it is the difference between the earliest and latest finish times for an activity. On the diagram it is simply the difference between the value in the head box of an activity arrow and the value in the lower right side of the event circle indicating the termination of the activity.
All those activities in which a zero shows in the total slack circle form the critical path. All other activities in the network have slack times, which means they can be delayed up to the value indicated without causing delay in project completion.
SCHEDULING
The third step in the management process, shown in Figure 1, is scheduling, the translation of a project plan (the previously prepared network diagram and the time calculations) into a timetable with calendar dates. These dates indicate the proposed start and completion time of individual activities and of the entire project.
The schedule for each activity is decided upon after considering its earliest and latest start and finish times previously calculated. The extreme limits for scheduling an activity are fixed by its earliest start time and its latest finish time. The activity can be scheduled to begin at any time between its earliest and latest start times. By beginning an activity after its earliest start time one must remember that the available slack time decreases and finally disappears if the latest start time is used. Of course, activities with no slack time have the same earliest and latest start time with no latitude for scheduling.
The results of an analysis such as that described above provide the basis for determining staffing, equipment and other resource needs and their scheduling. As project operations are carried out, evaluation of actual accomplishments should be continuously compared to the project plan and schedule. It will become immediately apparent if work is not proceeding according to schedule and the effects on future work can be predicted; and potential problems can also be detected and remedial action taken.
DAVIS, J.B. 1968 Why not PERT your next resource management problem? J. For. 66: 405-408.
HUSCH, B. 1969 A manual for the application of network analysis techniques to the planning, execution and control of UNDP projects. Rome, FAO.
