This information is part of the Business Simulation Library (BSL). Please support this work and ► donate.
The output y is the dimensionless degree of membership to the set of "good performances" or "good outcomes" according to a chosen type of membership function. Using the PerformanceIndicator allows to quickly grasp how a system is currently performing with regard to some criterion with a value of 1 indicating best possible performance and a value of 0 indicating worst possible (and possbily unsustainable) performance.
The following grid displays all available membership function types without inversion (invertResults = false
):
Membership Function Type | a | b | c | d |
---|---|---|---|---|
ramp | foot | shoulder | ||
triangular | left foot | center | right foot | |
trapezoidal | left foot | left shoulder | right shoulder | right foot |
sshaped | foot | shoulder | ||
sigmoidal | steepness of transition area | center of transition area | ||
pishaped | left foot | left shoulder | right shoulder | right foot |
psigmoidal | steepness of left transition | center of left transition area | steepness of right transition | center of right transition area |
gaussian | half width at half maximum | center of curve (maximum) | ||
bell | width of core | steepness | center of core |
When inversion is turned on (invertResult = true
) the result obtained from the membership function as specified will be subtracted from 1. Thus the ramp shown above can be made to ramp down from left to right with a
being the shoulder and b
the foot by inverting the results.
stockInfoInput
) will be normalized by divison either by a constant referenceValue (default = 1)
or by the exogenous input u (hasConstantReference = false
).+,-,±
) it is possible to indicate the basic relation between the amount in the stock and its performance value (e.g., +
indicating a "more-is-better" type of evaluation.