Forum Discussion
‘What-If’ X Start Date Constraint in the Microsoft Project Planner
Abstract
This paper reports the findings of an on-going PhD thesis in the University of Jos, Jos, Nigeria. The research evaluated and compared the effectiveness of the choice between ALAP and ASAP constraints in construction scheduling using the MS project software. A problem is identified from literature that users of the MS project are not well guided in the choice of start date constraints applied in scheduling project tasks. It is gleaned from literature that most practitioners who use the MS project software apply only as soon as possible (ASAP) constraint option which may not produce the best project outcome in every project scenario. Using ‘what-if’ questioning technique in a NON-EQUIVALENT quasi-experiment research design two hypotheses are developed and tested using ANOVA to evaluate and compare the effectiveness of choice between ALAP and ASAP constraints applied to schedule tasks in developing the construction schedules. The Null of hypothesis 1 states that: There is no significant relationship between the start date constraint and activity start variance. And the Null of hypothesis 2 states that: There is no significant relationship between the start date constraint and activity finish variance. Activity start variances were measured in a quasi-experiment for different scheduling options of ALAP and ASAP constraints and their differences tested using ANOVA if observed differences were due to mere chance or that they are due to different constraint types of ALAP and ASAP applied. Results show that FCAL = 81.05 and FTAB = 30.80 at 1% level of significance and FTAB = 9.55 at 5 % level of significance. Since FCAL is significantly large and FCAL > FTAB at both 1 % and 5 % levels of significance, we reject H0 and accept H1: That there is indeed a significant relationship between start date constraint and activity start variance at both 1 % and 5 % levels of significance. The probability of project completion (PR) was deductively determined as: PI = 100 – 65= 35%; PII = 100 – 59 = 41%; and PIII =100 – 57 = 43%. This data reduction shows clearly that the probability of project completion on the due date, PR improves with appropriate application of ALAP constraints to some selected tasks. The paper concludes that the new schedule alert feature which shows warnings and or suggestions when MS Project identifies possible scheduling conflict as well as the new task inspector tool which provides context-sensitive guidance for analyzing issues and taking corrective action, offering helpful and intuitive information about the task, such as factors that might affect the task schedule in MS Project 2010 and 2013 should be extended in deciding when to apply as soon as possible constraint or as late as possible constraint. The MS project and the Primavera project scheduling software developers are called upon to apply these new features to assist the users chose appropriately between the as soon as possible constraint or as late as possible constraint.
Key words
What-If, ALAP or ASAP, Start Date Constraint, Microsoft Project, ANOVA,
Discussion of Results of the Empirical ‘What-If’ Case Study
The decision rule in ANOVA is simple. It involves assessing and comparing two statistics, the F critical also called F calculated (FCAL) and the F read from table, FTAB, at stated levels of significance with defined denominator and numerator degrees of freedom (Spiegel and Larry, 2008). If FCAL is significantly large, and it is greater than FTAB, (FCAL ˃ FTAB), we can reject the null hypothesis H0, that there is no significant difference between the means and thus we can conclude that at the stated level of significance, the difference in activity start date is actually due to the constraint type applied in developing the initial baseline schedule. If however FCAL is less than 1 and or that FCAL ˂ FTAB we can conclude that there is no significant difference between the means and thus accept H0 and reject H1. In summary, in most cases if FCAL ˂ FTAB: H0 is accepted and H1 is rejected.
Conventionally used levels of significance are 0.1, 0.05 and 0.01. A 0.05 level of significance permits 5 percent probability for type 1 error. A decision made to reject a hypothesis at this level of significance has 5 percent chance of being wrong in rejecting a hypothesis that is possibly true that should not have been rejected. If rejected at 0.10 level of significance there is a 10 percent chance that the decision is wrong. If rejected at 0.01 level of significance there is only 1 percent chance that the decision is wrong, this supports Osuala (1990) assertion that the 0.01 level of significance is more exacting than the 0.05 and 0.10 levels of significance.
From the above analysis FCAL = 81.05 and FTAB = 30.8 at 1% level of significance and FTAB = 9.55 at 5 % level of significance. Since FCAL is significantly larger than 1, and it is greater than FTAB, (FCAL ˃ FTAB), we can reject the null hypothesis H0, that there is no significant difference between the means and thus we can conclude that at the stated level of significance, the difference in activity start date is indeed strongly correlated to the constraint type applied in developing the baseline schedule. The trend shown in Table 4 has not occurred by chance. Therefore it can be concluded that treatment III (Late start schedules) is a much more level headed and desirable schedule with reduced activity start variability which is shown to enhance schedule performance that events are likely to occur as planned.
This result shows the achievement of the third research objective which sought to test the hypothesized causal correlation or reasons for multivariate relationship of start date constraint and key performance indicators measured in terms of activity start variance. Though more field data is needed to validate this supposition, this study can at this stage assert that if the probability of project completion on the due date is inductively defined as: PR = 100 % - % VAR. Where % VAR = % activity start variance. A reduction of % VAR to ZERO would mean 100 % probability of project completion on the due date. If there is however a high % VAR, e.g, 75% VAR, then, the probability of project completion on the due date is only 25%. From Table 4 the average of VARI, VARII, VARIII = 175.5, 160.5, & 154 respectively. Therefore the percentage activity start variance for the different three scheduling treatment options of strong ASAP schedules; intermediate ALAP schedules and strong ALAP schedules are:
% VARI = 175.5/ 272 x 100 = 65 %
%VARII = 160.5 / 272 x 100 = 59 % and
% VARIII = 154/ 272 x 100 = 57 % respectively.
The probability of project completion (PR) on the due date for the different constraint options is estimated as: PI = 100 – 65= 35%; PII = 100 – 59 = 41%; and PIII =100 – 57 = 43%. Since application of 18 ALAP tasks improved the probability of project completion on the due date by 41 – 35 = 6%. And by applying 36 ALAP tasks, probability of project completion improved by 43 – 35 = 8%, we can conclude that:
% VAR varies as ASAP equation 2
% VAR varies as 1 equation 3
ALAP
Putting equations 2 and 3 into the earlier defined probability of project completion function:
PR = 100 – VAR % equation 1
The effectiveness of any schedule can be evaluated by assessing the relationship between the number of ASAP and ALAP tasks. As ALAP tasks increases, the %VAR decreases and as ALAP tasks decreases the %VAR increases and the effect of these decreases and increases in %VAR in equation 1 is improved probability of project completion in the case of decreased %VAR and reduced probability of project completion in the case of increased %VAR. Vividly:
(i) 0% of ALAP tasks yielded 65% of %VAR producing a 35% of PR
(ii) 7% of ALAP tasks yielded 59% of %VAR producing a 41% of PR and
(iii) 13% of ALAP tasks yielded 57% of %VAR producing a 43% of PR
This data reduction of the case study project shows very clearly that the probability of project completion on the due date, PR is greatly improved with application of ALAP constraints to selected tasks in developing the baseline schedule. These results emphasize the importance of scheduling to reduce activity start variance.
Implications of Findings for Future Research and Practice
The extent to which early start schedule, ASAP is or is not a practical and effective scheduling option in every scenarios being questionable has been confirmed in this study. A significant and strong relationship between start date constraints and activity start/finish variance is shown very clearly in the study. Therefore building and civil engineering contractors should schedule some project activities as late as possible while others which meet the criteria discussed in Efole (2009) are scheduled as early as possible. Actual occurrences of events will better properly align in desired planned order when some tasks, depending on their specific attributes like needing information request and information released, submittal approval and long lead supply items are scheduled to occur as late as possible. If this rule is followed as demonstrated in this research project events are likely to occur as planned which ensures more schedule reliability and schedule effectiveness. The software developers (Microsoft and Primavera) should extend some features like the new schedule alert tool and the new task inspector tool in the software architecture to assist users make appropriate choice of ALAP and ASAP constraint in different project scenarios.
Research Self-Assessment and the Remaining Work
All the research objectives have been addressed and achieved to a very large extent. This progress report shows how substantial findings have been achieved. The first objective which attempts to make a state of the arts review of context literature identified and justified the research problem. It is clear from the literature that there is gap in knowledge as well as a gap in practice regarding the appropriate choice of start date constraints in different project scenarios in the Microsoft Project Planner software. And this is well captured in the background information on constraints in the Microsoft project planner software discussed earlier. The results of the second and third objectives are presented in Tables 2 to 11. The distribution of the constraint types applied is shown in Tables 2 and 3. This finding is largely in agreement with literature that most practitioners apply only as soon as possible constraints which may not produce the best project outcome in every project scenario. The development of a framework to evaluate and compare the effectiveness between ALAP and ASAP has also been achieved and is presented below. This simple procedural framework inductively derived shows that:
% VAR varies as ASAP equation 2
% VAR varies as 1 equation 3
ALAP
This evaluation framework shows that application of ASAP constraints is directly proportional to the activity start variance and that application of ALAP constraint is inversely proportional to the activity start variance that may likely be experienced. Following this outstanding salient research findings a proposal is being drawn up to be sent to both the Microsoft incorporated and the Primavera project planner developers to help incorporate these findings in these pieces of scheduling software to improve their operation and usefulness.