Excavate to
Set Out
Strip Sele
Red Levels
1
Fig. 4.
Fuild Manholes,
Excavate for
Drains
Lay Drains
60
Cone Fdns
70
Start xC. to Fdas
Finish Exe
to Fans
Start Bwk to DP.C.
40
50
tion
Hardcore to GF. Slat
Hlinding to G.F Slab
Finish Brickwork to DPC
Fwk to Gr. Beams
Reint to Gr Brams
Conc. to
Gr. Beams
120
30
Start Conr to G.F. Slab
150]
Roof
Structure
20
Roof Coverings
Exti
Plumbing
Erect Frame
10
Fig. 5.
Glazing
Carpentry
Elec
Wiring
30
Brick work
Partitions
vities are linked with others and form a closed circuit.
Dangling This occurs when an arrow is left suspended with no clos- ing activity.
Numbering Events
Ext! Plumbing
Errors such as Looping and Dang- ling can be avoided if a rigid number- ing routine is adopted. A method suggested by Fulkerson1 is both simple and logical.
Fulkerson's Routine
(1) Find the initial event and num- ber it. An initial event has
arrows emerging from it but none entering.
(2) Delete all arrows emerging from numbered events. This will create new initial events.
(3) Number all new initial events. (4) Delete all arrows emerging from the newly numbered events creat- ing more initial events. Carry on this process until all events have been numbered.
Using this routine, it will be found impossible to create new initial events in a network containing a Loop. Dangling will become obvious by the inspection required when carrying out the numbering.
A numbering system which in- creases in increments of five or ten leaves more scope for the insertion and subsequent numbering of ad- ditional activities, e.g. 11, 12, 13, etc.
Glazing
Extl Painting
Plastering
Extl Painting
Assigning Duration
о
Intl Painting
Having drawn the diagram and completed the numbering, the next step is to ascribe durations to the acti- vities. This should be done by con- sultation between planner and site engineer. For many operations, the work contents will be well known and reasonably accurate estimates can be given. In exceptional cases, work study study synthetics may be available. These can be used to build up dura- tion times.
Where the duration cannot be estimated with any certainty, the
PERT method of three time estimates can be used. The engineer should be asked to give three separate estimates of the possible duration.
i
Optimistic (O). The time the acti- vity would take assuming every-
Reinforcement to Cols
(3)
thing goes as well as it possibly
can.
ii Most Likely (M). The time the ac- tivity would take in normal cir-
cumstances.
iii Pessimistic (P). The time the acti- vity would take if nearly every- thing goes wrong.
The weighted average is then cal- culated from the formula:-
0 + 4M + P =
6
expected time (Te)
The PERT formula is based on the assumption that estimates of the duration follow a Beta distribution. There is a measure of disagreement among mathematicians on the com- plete validity of the assumptions made. in the establishment of the PERT formula for a detailed discussion of the PERT statistics, reference should be made to papers by Grubbs
Clark and Fulkerson.2
Having estimated the durations, these should be marked on the diag- ram for analysis.
Analysing The Diagram
The
runs
The system of analysis sorts jobs into critical and non-critical. chain of critical jobs which through the network is the critical path. There will always be at least one critical path. The duration of the project is controlled by cumulative durations of individual jobs on the critical path.
Earliest Event Times
Before an event is reached, all the activities leading to that event must be completed. This is achieved by
Fmwk to cols
Conc Cols
70
80
(3)
(2)
90
Bwk to roof level
Concrete Slab
and Stairs
50
60
(2)
Cure and Strike
(10)
(4)
Reinforcement to
Kitchen area
Concrete to Kitchen area
(110)
(120)
(4)
(6)
Fig. 6.
Fmwk to Kitchen area
(7)
(100)
48
Far East Architect & Builder June, 1967