30 Material Criticality Assessment for Business Purposes …
465
The proposed fuzzy linguistic approach has the advantage to represent directly the
linguistic terms and preserve the imprecision within the representation modelled by
membership function. Since the risk metrics are modelled under the same scale, the
critical levels of all the metrics can be aggregated easily to obtain an overall critical
level, which also contains the imprecision. As a result, the critical levels of materials
are ranked allowing comparison. The capability of the ranking system enables the
decision makers to mitigate the business risks and maximize opportunities for raw
material selection.
This study also considers the importance of each metric to the overall criticality
level, so the importance weight is subjectively attributed to each risk metric in terms
of natural language, then modelled by the membership function. For example, the
price volatility risk is considered to be very important, because the price volatility
is calculated on the historic price fluctuation that may not predict the future price
variation. Hence, there is uncertainty in the price volatility risk that affects the assessment of the overall criticality. The integration of importance weight into risk metrics
allows the capture of experts’ knowledge regarding the sensitivity of risk metrics.
However, the users should be aware that importance weighting is likely to be subjectively different among the experts and it is sometimes difficult to reach a consensus.
Therefore, in a real case, the justification of weighting attribution should be well
documented or a specific workshop should be organized to create consensus among
the experts. For example, the Delphi method is commonly used for qualitative expert
elicitation, delivering group consensus (Helmer and Gordon 1968).
30.5 Conclusion
This study applies fuzzy linguistic approach to assess the criticality of raw materials with multiple risk metrics when quantitative data are limited or non-existent.
As the material criticality is determined by various factors that usually change over
time, so the indicators of risk metrics should also be temporally updated according
to the specific business perspectives. More indicators are considered, more data are
required. Since detailed quantitative data capture requires significantly more time
and economic resources, linguistic information is relatively easy to be captured.
This approach models appropriately linguistic information with imprecision, which
expresses the experts’ knowledge in a user-friendly format - natural language.
However, the use of linguistic terms loses the precision comparing with the numerical
values. Therefore, the users should be aware whether or not their underlying studies
require high level precision. From our perspective, this research could extend to
more strategic functions, for example, integrating proactive judgments with the risk
metrics may help the decision makers to identify potential future critical materials.
465
The proposed fuzzy linguistic approach has the advantage to represent directly the
linguistic terms and preserve the imprecision within the representation modelled by
membership function. Since the risk metrics are modelled under the same scale, the
critical levels of all the metrics can be aggregated easily to obtain an overall critical
level, which also contains the imprecision. As a result, the critical levels of materials
are ranked allowing comparison. The capability of the ranking system enables the
decision makers to mitigate the business risks and maximize opportunities for raw
material selection.
This study also considers the importance of each metric to the overall criticality
level, so the importance weight is subjectively attributed to each risk metric in terms
of natural language, then modelled by the membership function. For example, the
price volatility risk is considered to be very important, because the price volatility
is calculated on the historic price fluctuation that may not predict the future price
variation. Hence, there is uncertainty in the price volatility risk that affects the assessment of the overall criticality. The integration of importance weight into risk metrics
allows the capture of experts’ knowledge regarding the sensitivity of risk metrics.
However, the users should be aware that importance weighting is likely to be subjectively different among the experts and it is sometimes difficult to reach a consensus.
Therefore, in a real case, the justification of weighting attribution should be well
documented or a specific workshop should be organized to create consensus among
the experts. For example, the Delphi method is commonly used for qualitative expert
elicitation, delivering group consensus (Helmer and Gordon 1968).
30.5 Conclusion
This study applies fuzzy linguistic approach to assess the criticality of raw materials with multiple risk metrics when quantitative data are limited or non-existent.
As the material criticality is determined by various factors that usually change over
time, so the indicators of risk metrics should also be temporally updated according
to the specific business perspectives. More indicators are considered, more data are
required. Since detailed quantitative data capture requires significantly more time
and economic resources, linguistic information is relatively easy to be captured.
This approach models appropriately linguistic information with imprecision, which
expresses the experts’ knowledge in a user-friendly format - natural language.
However, the use of linguistic terms loses the precision comparing with the numerical
values. Therefore, the users should be aware whether or not their underlying studies
require high level precision. From our perspective, this research could extend to
more strategic functions, for example, integrating proactive judgments with the risk
metrics may help the decision makers to identify potential future critical materials.