Decision Making in Systems Engineering and Management, 3rd Edition

Appendix ARealization Analysis Levels 0 and 2

The detail contained in this appendix is and extension of the Level 1 application discussed in Section 10.7.1. It is an adaptation of the authors' work [1] is presented here for completeness regarding realization analysis. As noted in Chapter 10, Level 0 and Level 2 realization analysis steps are used to address specific concerns that are not in common to all systems decision support projects involving uncertainty. Consequently, they appear here rather than in the primary Chapter 10. The figure below is repeated from the chapter for convenience.

A.1 Level 0 Analysis—Refined Choice Set Identification

Consider the Monte Carlo simulation output for the six alternatives shown in the righthand image of Figure A.1, which displays realizations of each alternative and significant overlap between the cloudplots. This notion of “overlap” is broader than the obvious commingling of realizations between alternatives. In Figure A.1, even though Alt B's realizations do not intertwine with Alt A's, several of Alt A's least cost realizations are cheaper than Alt B's most expensive ones.

Similarly, despite Alt B's generally higher value, several of Alt A's best‐performing realizations produce greater value than Alt B's least valuable ones. In both cases, there is a non‐zero probability that a randomly selected pair of realizations from the Alt B ...