Page 202 - Materials Chemistry, Second Edition
P. 202

L1644_C05.fm  Page 175  Monday, October 20, 2003  12:02 PM









                                       The term simulation can be understood as an analytical method meant to imitate
                                    a real-life system, especially when other analyses are too mathematically complex
                                    or too difficult to reproduce. Without the aid of simulation, a spreadsheet model
                                    would only reveal a single outcome, generally the most likely or average scenario.
                                    Spreadsheet uncertainty analysis uses a spreadsheet model and simulation to analyze
                                    the effect of varying inputs or outputs of the modeled system automatically. The
                                    random behavior of how MC simulation selects variable values to simulate a model
                                    is similar to that employed by games of chance. When a player rolls a die, he or
                                    she knows that a 1, 2, 3, 4, 5, or 6 will come up, but does not know which will
                                    occur in any particular roll. It is the same with the variables that have a known range
                                    of values but an uncertain value for any particular time or event (Decisioneering
                                    1996).


                                    5.4.1  USE OF SOFTWARE PACKAGES TO PROPAGATE UNCERTAINTY
                                          BY PERFORMING MONTE CARLO SIMULATION

                                    In order to simplify the task of determining the uncertainty of a parameter by MC
                                    simulation, various commercial software packages are available. Among them, Crys-
                                                                                        ‚
                                                                           ‚
                                                   ‚
                                          ‚
                                    tal Ball ,* @Risk ,**  Analytica,*** Stella II ,**** PRISM ***** and Susa-
                                      ‚
                                    PC ****** can be highlighted. Table 5.1 summarizes the main information about
                                    each of these software packages.
                                       According to Metzger et al. (1998), @Risk was originally designed for business
                                    applications, but it has found wide use in human health and ecological risk assess-
                                    ment. This package includes uncertainty in estimates to generate results that show
                                    all possible outcomes. Because of its standardized spreadsheet backbone, @Risk is
                                    easy to use without a need for extensive statistical knowledge, modeling capability
                                    or programming ability.
                                       Analytica and Stella II are two stand-alone programs designed for a wide variety
                                    of applications. Analytica is a model-building program that attempts to simplify
                                    sophisticated systems with the use of multilevel influence diagrams. Stella II is a
                                    multilevel hierarchical environment for constructing and interacting with models.
                                    These two programs are designed to simplify complex problems and therefore
                                    require additional effort in learning the software and building the models. However,
                                    because these models are displayed in diagrams, their interpretation for other users
                                    tends to be easier than for any other software package.
                                       PRISM and Susa-PC are Fortran-based codes, designed specifically for use in
                                    risk analysis. PRISM is a simple free-access program that builds distributions for
                                    input into any model and then analyzes the output of that model. Susa-PC is a more


                                    * Registered trademark of Decisioneering, Inc., Denver, CO.
                                    ** Registered trademark by Palisade Corporation, Newfield, NY.
                                    *** Registered trademark of Decisioneering, Inc., Denver, CO.
                                    **** Registered trademark of High Performance Systems, Inc., Lebanon, NH.
                                    ***** Registered trademark of SENES Oak Ridge, Inc., Oak Ridge, TN.
                                    ****** Registered trademark of Gesellschaft für  Anlagenund Reaktorsicherheit (GRS) mbH, Köln,
                                    Deutschland.

                                    © 2004 CRC Press LLC
   197   198   199   200   201   202   203   204   205   206   207