Because heuristics are designed to
make a trade-off between effort and
accuracy (Gigerenzer & Gaissmaier,
2011), complex situations call for
either more advanced solution methods or for an increase in additional
resources and managerial attention,
as established by Shenhar (2001).
2. The effect of uncertainty greatly depends
on its impact.
Uncertainty was defined as a variation in the duration of an activity.
Because most variations led to activity delays, a higher degree of uncertainty generally led to longer project
duration; therefore, a high degree of
uncertainty, combined with a severe
penalty for deadline overruns, led to
steep cost increases.
3. Individuals vary in how they assess
complexity and uncertainty.
This variation can be partly due to
the experience or capability of the
project managers in assessing those
contextual factors. In the experiment, this was addressed by means
of a threshold function, in which the
actual and perceived levels of complexity and uncertainty were varied.
A discrepancy between the actual
and perceived levels of complexity or
uncertainty led to a judgment error.
We came to the conclusion that the
direction of judgment errors is crucial.
Perceiving a project as highly complex
and uncertain, although not true in
reality, yields significant advantages
compared with the opposite scenario. We recommend project managers who are incapable of correctly
assessing a project’s complexity and/
or uncertainty (e.g., through limited
information) to err on the safe side.
4. We identified the conditions in which
each solution strategy thrives.
The time-based solution strategy per-
forms particularly well when deadline
selecting the 25th and 75th percentiles,
respectively. Both solution strategies
share a decreasing deadline deviation
trend as the deadline increases, with-
out leading to different conclusions for
the overall performance. The SP factor
was varied from 0.1 to 0.9 with steps of
0.1. No consistent trend for the solution strategies across the complexity
and uncertainty dimensions could be
Discussion and Conclusion
Three contributions have been made in
this article. First, the decisions of students throughout the Project Scheduling
Game were translated into two major
solution strategies; these are comprised
of five building blocks, namely focus,
activity criticality, ranking, intensity, and
action. The first solution strategy focuses
on time and employs three mechanisms
to approach the deadline. The Greatest Rank Positional weight priority rule
is used, as well as a buffer based on
the slack duration ratio of Hazir et al.
(2010), and a final check to protect the
deadline is performed. The second solution strategy focuses heavily on costs, at
the expense of an increased exposure to
risk. The average most expensive priority rule is used to rank activities. Elitism
is applied only to accept cost improvements, and non-critical activities’ slack
is consumed to a larger degree.
Second, complexity and uncertainty
were included as contextual factors.
The literature overview in the Introduction indicated that these are dominant
themes and that a link between complexity and project outcome (Hanisch
& Wald, 2011) and a continued study
of uncertainty (Hall, 2012) were among
the challenges for future research. To
that end, we have conducted a large
computational experiment that allows
us to quantify the impact of complexity
(Maylor et al., 2008) and uncertainty.
The following five conclusions can be
drawn from this experiment:
1. A high degree of complexity has a
negative effect on the cost deviation.
using focus and intensity. The focus and
intensity settings for the baseline scenario
were described in the section, “Proposed
Strategies.” In this section, the effect of an
increased level of effort is studied. Three
separate experiments were conducted to
study the effect of an increased level of
effort on the performance of the solution
strategies. The global cost deviations of
the three experiments can be found in
Table A1 of the Appendix. The main findings for each experiment can be summarized as follows:
1. Experiment 1 adopted a focus of 100%
in absence of any uncertainty (U 0).
The intensity was varied from 0.6
to 1.0 in steps of 0.1. The results
indicate that an increased intensity
leads to better cost deviations.
2. Experiment 2 adopted an intensity of
100% in absence of any uncertainty
The focus was varied from 0.6 to 1.0
in steps of 0.1. Similar to the first
experiment, the global cost deviation decreased as the focus was
increased, but the decrease was less
steep compared with the findings of
the first experiment.
3. Experiment 3 reintroduced the uncertainty settings of the baseline scenario, whereas the focus was kept at
100% and the intensity was varied
again from 0.6 to 1.0.
Hence, compared with the first
experiment, these settings allowed us
to explore the influence of the uncertainty. As the intensity (and thus the
level of effort) increased, the global
cost deviation decreased; however,
the cost deviations are higher than
those of the first experiment, which
can be attributed to uncertainty
affecting the activity durations.
Finally, we also tested the influence of the deadline and the SP level.
The deadline parameter was varied by