sensitivity: Sensitivity matrix calculator

In gripp: General Inverse Problem Platform

Description

This sotfware will calculate the sensitivity matrix for the Direct Problem. First order derivatives are calculated using central difference approximation.

Usage

sensitivity(parm_s)

Arguments

parm_s

Set of values to be considered as parameters for the Direct Problem solution

Value

A matrix with the derivative of the function that represents the Direct Problem for each parameter.

Examples

#
# Configure the name of the folder where each file with the parameter to be changed.
# If the folder is 'wd' then it will not be changed.
# You can change it to another location where your parameter files are.
auxi <- system.file(package = "gripp")
folder_name <- c(auxi,auxi)
#
# Configure the name of each file to be changed in the Direct Problem Solver.
file_name <- c('f1.R','f1.R')
#
# Configure the name of each parameter to be changed.
parm_name <- c('A','B')
#
# Configure alternative method using line number to enter each parameter
# When line_number is zero, then the input method uses keywords.
line_number <-c(0,0)
#
# Configure each parameter type, where:
#   1 means a numeric variable informed as a string, such as parm <- "1.38"
#   2 means a numeric vector informed as a string of numbers separated with a space character.
#     For example, parm <- "1.25 3.4". You must tell which one will be the variable considered.
parm_type <- c(1,1)
#
# Informe the position of the variable to be considered in the problem in the vector.
# If the parameter is a numeric variable, then its position is zero.
parm_vector <- c(0,0)
#
# Configure the smallest value for each parameter
parm_min <- c(0,0)
#
# Configure the larger value for each parameter
parm_max <- c(2,5)
#
# Configure the name of the folder where the command must be called.
# If the folder is 'wd' then it will not be changed.
# You can change it to another location where your parameter files are.
command_folder <- auxi
#
# This variable means that the Direct Problem is solved using R or outside
# if this is FALSE, the results are to be read from a file.
# When it is TRUE, results will be passed inside R using the variable "results"
isitR <- TRUE
#
# Configure the command to be used to call the Direct Problem Solver.
# if this is FALSE, the results are to be read from a file.
# When it is TRUE, results will be passed inside R using the variable "result"
command <- 'f1.R'
#
# Parameter positive and negative percentual difference to be used to calculate the derivative
# ppdif must me a number between 0 and 100
# parameter_pos <- parm + (ppdif/100)*(parm_max-parm_min)
# parameter_neg <- parm - (ppdif/100)*(parm_max-parm_min)
ppdif <- 1
#
# Configure a string to be used to attrib values in the Direct Problem file.
attrib_str <- '<-'
#
# Configure the name of the file with the results obtained by the Direct Problem.
# It must be a single column of values.
# The results are changed at each run by the Direct Problem Solver.
result <- 'result.dat'
#
# Configure the name of the folder where the results can be found after each run.
# If the folder is 'wd' then it will not be changed.
# You can change it to another location where your result file is.
result_folder <- tempdir()
#
sensitivity(c(1,3))

gripp documentation built on Aug. 24, 2019, 9:03 a.m.