Sensitivity Analysis - Morris method (advanced)

Application of the Morris method with a reduced risk of factors underestimation

Sensitivity analysis is used to estimate the influence of uncertainty factors on the output of a function. The Morris method is sometimes referenced to as a qualitative method : it gives rough estimations with a limited number of calculations. The Morris method can be used to simplify a function, as a first step. It can identify the factors with a low influence which can be fixed. For further information :
Saltelli, A., Tarantola, S., Campolongo, F., and Ratto, M. (2004). Sensitivity Analysis in Practice - A Guide to Assessing Scientific Models. Wiley.

This algorithm reduces the risk to underestimate and fix non-negligible factors. It is presented in:
Henri Sohier, Helene Piet-Lahanier, Jean-Loup Farges, Analysis and optimization of an air-launch-to-orbit separation, Acta Astronautica, Volume 108, March–April 2015, Pages 18-29, ISSN 0094-5765, http://dx.doi.org/10.1016/j.actaastro.2014.11.043.

For the classic Morris method : http://fr.mathworks.com/matlabcentral/fileexchange/48883-sensitivity-analysis-morris-method--simple-

IMPORTANT : The function analyzed must be a function with one multidimensional input x. x must represent the values of the uncertainty factors in the quantiles hyperspace (the i-th coordinate of x is not the actual value of the i-th factor, but the corresponding value of its cumulative distribution function). To adapt your function, first calculate the actual values of the factors by applying the inverse of their cumulative distribution function to x; Matlab includes such inverses: mathworks.com/help/stats/icdf.html ).