If you wish to let values of R vary from run to run along a normal distribution
with known mean and standard deviation, choose, within “Sensi Specs” the “Distribution” option instead of “Incremental”. When you specify “Seed” as a positive
integer, you ensure that the model will replicate a particular random number
sequence in subsequent sensitivity runs. Specify 0 as the seed and a “random”
seed will be selected (Fig. 2.8). If you do not wish to make use of the normal
distribution of the random numbers used in the sensitivity analysis, click on the bell
curve button. This curve bell-shaped button will change its appearance when you
click on it, and then allow you to specify a minimum, maximum and a seed for your
sensitivity analysis.
Another choice for the specification of sensitivity runs in STELLA is not to
change parameters in incremental intervals or along distributions. You can specify
ad hoc values for each of the consecutive runs.
You can easily specify a whole series of parameter variations in STELLA and
make the hundreds of runs needed to reasonably explore the combinations for their
collected sensitivity. Examination of the results can lead you to those parameter
groups that can cause trouble. You might consider eliminating such a combination
by a structural change in the model or by investing more effort in narrowing the real
range of these parameters through extended research.
Fig. 2.8
36
2 Exploring Dynamic Biological Systems
with known mean and standard deviation, choose, within “Sensi Specs” the “Distribution” option instead of “Incremental”. When you specify “Seed” as a positive
integer, you ensure that the model will replicate a particular random number
sequence in subsequent sensitivity runs. Specify 0 as the seed and a “random”
seed will be selected (Fig. 2.8). If you do not wish to make use of the normal
distribution of the random numbers used in the sensitivity analysis, click on the bell
curve button. This curve bell-shaped button will change its appearance when you
click on it, and then allow you to specify a minimum, maximum and a seed for your
sensitivity analysis.
Another choice for the specification of sensitivity runs in STELLA is not to
change parameters in incremental intervals or along distributions. You can specify
ad hoc values for each of the consecutive runs.
You can easily specify a whole series of parameter variations in STELLA and
make the hundreds of runs needed to reasonably explore the combinations for their
collected sensitivity. Examination of the results can lead you to those parameter
groups that can cause trouble. You might consider eliminating such a combination
by a structural change in the model or by investing more effort in narrowing the real
range of these parameters through extended research.
Fig. 2.8
36
2 Exploring Dynamic Biological Systems