Some Basic Considerations on the Design and the Interpretation of Indicators. . .
5
This use is based on two basic ideas: firstly, the fact that the indicator value serves
as an indicator for a more or less complex and less transparent system state, and
secondly, that several influencing variables can be combined in such an indicator
value, which, as an aggregated value, provide indications of the current behavior of
the overall system.
Typical examples are warnings regarding the condition of complex industrial
plants or in intensive care medicine to summarize the values from various vital
parameters.
2.1.2 Decisions Between Alternatives
While in the first case the scale of the indicator together with an absolute threshold
value comes to the fore and requires special design considerations, the second
field of application requires an indicator design that evaluates different decision
alternatives and thus allows a comparison of these alternatives. Typical examples are
the classic advantage and disadvantage lists for previously given decision scenarios
and a decision for the overall problem derived from the individual arguments
collected in these lists. In this case, the focus is not so much on the state of the
system itself, but much more on a relative evaluation, a ranking that relates the
different scenarios of the decision problem to one another.
2.1.3 Optimization
A much more complex use of indicators is found in the solution of system
optimization tasks. In this case it is assumed that the behavior of a system
can be influenced by setting parameters (the so-called manipulated variables),
whereupon the value of the target variable changes. Through targeted changes of
the manipulated variables, an optimum of the target function value is to be achieved
iteratively during optimization.
The description of the optimization procedure clearly shows the use of indicators:
The indicator fulfils the function of the so-called target function and thus summarizes the system state achieved by setting the manipulated variables on a single scale.
On this scale, the mechanism of the optimization algorithm then takes effect and
iteratively minimizes or maximizes the target function or indicator value.
2.1.4 Modelling Real Systems
Similar to the use for triggering warnings, several indicators can be recorded and
observed simultaneously and their dynamic change in indicator values can be
interpreted as an image of the underlying real system. Once again, the example
from intensive care medicine is the most vivid: the measured vital parameters are
not combined into one indicator variable, but their value progression is visualized
5
This use is based on two basic ideas: firstly, the fact that the indicator value serves
as an indicator for a more or less complex and less transparent system state, and
secondly, that several influencing variables can be combined in such an indicator
value, which, as an aggregated value, provide indications of the current behavior of
the overall system.
Typical examples are warnings regarding the condition of complex industrial
plants or in intensive care medicine to summarize the values from various vital
parameters.
2.1.2 Decisions Between Alternatives
While in the first case the scale of the indicator together with an absolute threshold
value comes to the fore and requires special design considerations, the second
field of application requires an indicator design that evaluates different decision
alternatives and thus allows a comparison of these alternatives. Typical examples are
the classic advantage and disadvantage lists for previously given decision scenarios
and a decision for the overall problem derived from the individual arguments
collected in these lists. In this case, the focus is not so much on the state of the
system itself, but much more on a relative evaluation, a ranking that relates the
different scenarios of the decision problem to one another.
2.1.3 Optimization
A much more complex use of indicators is found in the solution of system
optimization tasks. In this case it is assumed that the behavior of a system
can be influenced by setting parameters (the so-called manipulated variables),
whereupon the value of the target variable changes. Through targeted changes of
the manipulated variables, an optimum of the target function value is to be achieved
iteratively during optimization.
The description of the optimization procedure clearly shows the use of indicators:
The indicator fulfils the function of the so-called target function and thus summarizes the system state achieved by setting the manipulated variables on a single scale.
On this scale, the mechanism of the optimization algorithm then takes effect and
iteratively minimizes or maximizes the target function or indicator value.
2.1.4 Modelling Real Systems
Similar to the use for triggering warnings, several indicators can be recorded and
observed simultaneously and their dynamic change in indicator values can be
interpreted as an image of the underlying real system. Once again, the example
from intensive care medicine is the most vivid: the measured vital parameters are
not combined into one indicator variable, but their value progression is visualized